WERF Endometriosis Phenome and Biobanking Harmonisation Project for Experimental Models in Endometriosis Research (EPHect-EM-Homologous): homologous rodent models

Endometriosis affects an estimated 190 million women (and those assigned female at birth) worldwide with a significant personal and societal burden due to its painful and fertility-related symptoms ( Nnoaham et al. , 2011 ; Zondervan et al. , 2020 ). There is no known cure, and treatments are associated with low long-term success rates and significant side effects ( Johnson et al. , 2013 ). Endometriosis is defined as endometrial-like tissue outside the uterus; however, this definition does not encompass the complex symptomatic, pathobiological, and multisystemic nature of the disease ( Zondervan et al. , 2020 ). The overwhelming prevalence of endometriosis and lack of known aetiology highlights the need for a better understanding of the basic biology of endometriosis formation and persistence to enable subsequent identification of targets for effective therapies for endometriosis. Endometriosis lesions are complex multicellular tissue deposits composed of endometrial-like stromal cells, epithelial glands, and extracellular matrix (ECM) deposition including fibrosis and scarring, sometimes with evidence of haemorrhage ( Clement, 2007 ; Vigano et al. , 2018 ). They are infiltrated by blood vessels and nerve fibres as well as abundant immune cells, yielding a highly inflammatory microenvironment ( Burns et al. , 2012 , 2018 ; Greaves et al. , 2017a ; Forster et al. , 2019 ; Hogg et al. , 2020 ; Hogg et al. , 2021 ; Panir et al. , 2022 ; Fattori et al. , 2024 ). The local cellular niche of a lesion is regulated by oestrogen, which has significant impacts on endometrial cell types ( Saunders and Horne, 2021 ) as well as processes such as neuroangiogenesis and neuroinflammation ( Greaves et al ., 2014a , 2015 ). The histological appearance of lesions varies significantly, with lesions having different contributions of ECM, fibrosis ( Li et al. , 2022 ) and endometrial-like cells ( Clement, 2007 ) and differences in their synchronicity with the menstrual cycle ( Colgrave et al. , 2020 ) as well as the extent of immune and nerve infiltration ( Tran et al. , 2009 ). It is likely that disease processes vary in each lesion, as suggested by the extreme variation in disease lesion subtype, symptom presentation, and response to treatment, as well as the hypothesized ‘life-cycle’ of lesions that progress through different stages and colours ( Zondervan et al. , 2020 ). Lesions are currently classified into three major subtypes: (i) superficial peritoneal, (ii) ovarian endometriosis (endometriomas), and (iii) deep lesions ( International Working Group of AAGL, ESGE, ESHRE and WES et al. , 2021 ). Associated genomic loci vary by lesion subtype, with most of the 42 loci identified to date associated with endometriomas ( Rahmioglu et al. , 2023 ). Researchers now understand more about the pathobiology of endometriosis than ever before; however, the exact aetiology of the disease remains elusive and may vary between patients and lesion subtypes. As with most conditions with known genetic and phenotypic complexity, current evidence confirms that there is unlikely to be a unified theory for endometriosis that explains the different lesion subtypes while also integrating epigenetic, genetic, and immune aberrations and postulated environmental exposures that are thought to contribute to the disease ( Zondervan et al. , 2018 ; Giudice et al. , 2023 ). Instead, the different lesion subtypes may represent different disease entities, with possible divergent origins ( Nirgianakis et al. , 2020 ; Imperiale et al. , 2023 ; Rahmioglu et al. , 2023 ). The most commonly invoked theory of aetiology is Sampson’s theory of retrograde menstruation, which states that menstrual effluent is disseminated into the peritoneal cavity via the fallopian tubes, where it migrates, attaches, grows, and invades peritoneal tissues as endometriotic lesions ( Sampson, 1927a , b ). However, Sampson’s theory is insufficient, as only ∼10% of reproductive-age women suffer from endometriosis ( Halme et al. , 1984 ), while most women, both with and without endometriosis, exhibit some levels of retrograde menstruation ( Halme et al. , 1984 ; Bartosik et al. , 1986 ; Kruitwagen et al. , 1991 ; Sharpe-Timms, 2005 ; O et al. , 2017 ). Nonetheless, a uterine origin provides a simple and attractive source for the cell types involved, and somatic genetic data may favour this aetiology ( Anglesio et al. , 2017 ; Suda et al. , 2018 ; Lac et al. , 2019a ; Moore et al. , 2020 ; Praetorius et al. , 2022 ; Yamaguchi et al. , 2022 ). Indeed, there are examples of clonality between uterine epithelium and endometriotic lesion epithelium ( Inoue et al. , 2019 ; Bulun et al. , 2023 ). Additionally, these mutations may contribute to fibrogenesis in lesions ( Guo, 2018a ) that selects for clones bearing these mutations. Other theories encompassing coelomic metaplasia, vascular and lymphatic spread, stem cells (and neonatal bleeding), Müllerian remnants ( Vigano et al. , 2018 ), and a combination of theories may explain at least some of the different forms/occurrences of endometriotic lesions ( Tsamantioti and Mahdy, 2022 ). Other contributing factors that have been identified recently are not restricted to any one theory of origin and include heritable features, immune microenvironment, somatic mutations, as well as insights provided through epigenomic, metabolomic and proteomic studies, as well as demographic, anthropometric, and/or other personal characteristics. Contributing further to the complexity and enigma of the disease, lesion phenotype and burden do not correlate with clinical symptoms of endometriosis ( Litson et al. , 2022 ). Previously, the World Endometriosis Research Foundation (WERF) Endometriosis Phenome and Biobanking Harmonisation Project (EPHect) established standard recommendations and minimum requirements for the collection of clinical and surgical data from those with endometriosis ( Becker et al. , 2014 ; Vitonis et al. , 2014 ), as well as a standardized physical examination assessment in endometriosis ( Lin et al. , 2024 ). WERF also developed internationally agreed-upon standard operating procedures (SOPs) for the collection, processing, and storage of tissue and fluid biospecimens ( Becker et al. , 2014 ; Fassbender et al. , 2014 ; Rahmioglu et al. , 2014 ; Vitonis et al. , 2014 ). Through this process, WERF identified an unmet need to develop standards of practice in experimental models of endometriosis. Rodent models of endometriosis are most commonly used for in vivo studies. These models are an attractive choice due to the short generation time, relatively low husbandry costs and space requirements, availability of genetically modified strains of mice and rats, and the ability to have large experimental groups ( Burns et al. , 2021 ). The first rodent model for endometriosis, describing auto-transplantation of homologous uterine tissues in the rat, was published in 1985 ( Vernon and Wilson, 1985 ). This model, with moderate modifications, has been used widely for studies on the mechanisms of attachment and growth of ectopic tissue and the pathophysiology of endometriosis, including pain and subfertility. Subsequently, one of the first mouse models of endometriosis aimed to replicate the rat model of endometriosis ( Cummings and Metcalf, 1995 ); however, in mice, multiple adaptations have emerged with the aim of improving physiological relevance and refining animal welfare in line with the 3Rs: Replacement, Reduction, and Refinement ( Hubrecht and Carter, 2019 ). Current homologous rodent models that are used to study endometriosis include both rats and mice and range from the surgical engraftment of uterine tissue to injection of ‘menses’-like endometrium ( Cummings and Metcalf, 1995 ; Grummer, 2006 ; Pelch et al. , 2010 ; Pelch et al. , 2012 ; Wilson et al. , 2020 ; Burns et al. , 2021 ). While each model has strengths and weaknesses and each has served a purpose in helping to understand the pathophysiology of endometriosis, it is important to consider each strength and weakness prior to the design of preclinical studies. In vivo models are required to determine the molecular underpinnings of endometriosis, to develop and test potential therapeutics, and to discover non-invasive biomarkers for disease diagnosis. However, the variety of preclinical models of endometriosis in use causes problems with replicability and direct comparison of study findings. An important concern for all animal research intended for the study of human disease is the translatability of the animal model and whether improvements in preclinical study design can enhance its relevance to human disease ( de Caestecker et al. , 2015 ). The process of drug development is complex, lengthy, and expensive; thus, the validity, scientific rigour, and predictive nature of the animal model used are paramount for successful therapeutic development ( Groothuis and Guo, 2018 ). Clinical trials testing potential therapeutics for endometriosis often fail due to lack of efficacy ( Guo and Groothuis, 2018 ), presumably because preclinical testing was not suitably robust or the translatability of the model was limited. Unfortunately, failed or informatively null studies are often not reported or published, even though these findings would inform the field and yield improved understanding of protocols, model validity, and treatments/drugs explored, and would prevent replication of studies that are unlikely to be successful. A possible concern for translatability in endometriosis research is the recognition that no single rodent model of endometriosis will fully recapitulate the disease among all endometriosis phenotypes for the following reasons; (i) In contrast to humans and some non-human primates, rodents do not develop endometriosis spontaneously (rodents remodel their endometrial lining during the oestrous cycle without endometrial shedding; a notable exception is the spiny mouse ( Bellofiore et al. , 2017 )); (ii) to date, no rodent model of experimental endometriosis reproduces all three main subtypes of endometriotic lesions (though not all humans experience all three subtypes); (iii) there is limited evidence on the true aetiology of endometriosis in women; and (iv) since rodents do not menstruate, the induced lesions likely do not undergo the repeated tissue injury and repair (ReTIAR), as proposed in humans – a process that facilitates lesion progression and fibrogenesis ( Zhang et al. , 2016a , b ; Guo, 2018b ). As such, the subsequent model may not recapitulate all the pathways involved in lesion development and progression in humans, nor can it accurately model all details of human endometriosis pathophysiology, such as endometriosis-associated dysmenorrhoea or subfertility. Acknowledging these hurdles, a working group was established to develop baseline standards for rodent models in endometriosis research and to develop internationally agreed-upon SOPs for experimental models in endometriosis to limit procedural variation and improve reporting standards to maximize inference and comparability without stifling scientific creativity. We also identified gaps in knowledge and offered recommendations for model design and reporting standards, which will allow researchers to compare across studies and laboratories to increase study fidelity, enhance the robustness of preclinical research, and improve the relevance of rodent models to human disease. Since the complex aetiology of endometriosis is not yet fully understood, our recommendations have endeavoured to ensure the model(s)/SOPs presented maintain sufficient flexibility to not hamper discovery, but to be used to help uncover novel and confirm known aetiology, pathogenesis, and disease progression. In endometriosis research, the main variables in rodent model design are as follows: (i) species and animal strain, (ii) location of ectopic tissue placement, (iii) volume of donor tissue, (iv) method of induction, (v) hormonal status of recipient, and (vi) donor uterine tissue ‘type’. Due to these variations and the increasing number of models and researchers entering the field, WERF initiated this project to develop recommendations for homologous rodent models for endometriosis research. Recommendations and SOPs were also developed for heterologous rodent models, endometriosis organoids, and pain behaviour, and the results are presented in the associated WERF companion papers ( Hull et al. , 2025 ; Marr et al. , 2025 ; and Dodds et al. , 2025 , respectively).

To analyse and assess existing rodent models for their ability to demonstrate the key hallmarks of endometriosis and standardize procedures, including visualization of lesions, pain-associated behavioural measurements, and/or investigation of the subfertility of the animals, internationally renowned endometriosis researchers were invited to join a consortium. Consortium members were identified based on pioneering work with specific rodent model(s) for endometriosis and the criteria of having published manuscripts with (i) evidence of comprehensive histological analysis, (ii) exemplary experimental design, and/or (iii) having published at least three research studies using the same model. For identification of publications, the search criteria were as follows: ‘endometriosis’ + ‘mouse’ + ‘murine’ + ‘rat’. Of the experts invited, 39 accepted. An additional six observers who were students or early career endometriosis researchers were also included. Of the 39 original working group members, 27 and 29, respectively, attended the workshops in June 2021. Eight working group members were added at a later date, and three left the group, leaving a total of 44 contributors. The aims of the workshops were to reappraise the current and emerging models of endometriosis and to develop protocols and criteria (see Supplementary Fig. S1 ) that meet a minimum set of standards based on the objective(s) of the study. All models were examined with the understanding that an individual model is unlikely to address all parameters of endometriosis, and instead the focus was on identifying research standards critical for the use of rodent models in the study of endometriosis. Following the workshops, consortium members were assigned or self-selected to one or more working groups based on their affinity and expertise ( Supplementary Fig. S1 ). The four groups were: (i) heterologous models, (ii) homologous models, (iii) models to assess pain behaviour, and (iv) organoids as an emerging ( in vitro ) model for endometriosis. For each group, a discussion forum was used to submit SOPs for circulation, standardization, and to enable a platform for discussions. Further discussions were held between authors to continue dialogues and develop consensus on areas for unification. For the homologous working group, 14 protocols were received from members of the group; these reflected the full spectrum of models described in the literature. A standard for how to format individual protocols was developed. Each investigator added additional requirements to their respective protocol(s) to ensure maximum detail for comparison across all protocols. Next, the protocols were examined for similarities, differences and key areas that could lead to new scientific discovery.

The consortium defined lesions in rodent models of experimental endometriosis as the establishment of uterine tissue outside the uterine cavity. Acknowledging the caveat that rodent models cannot fully recapitulate human disease, our focus was on the existence of key endometriosis-like lesion characteristics in the models. Endometriosis is defined as immune-infiltrated ectopic endometrial-like tissue, consisting of stromal cells and glandular epithelial cells, the presence of hemosiderin, and an accumulation of ECM consistent with fibrosis ( Hsu et al. , 2010 ; Khan et al. , 2014 ; Vigano et al. , 2018 ; Burney, 2022 ). For the basis of any current or future model, endometriosis-like lesions need to meet these criteria ( Fig. 1 ). We recognize that lesion histology is incredibly diverse and that not all lesions will contain all of the described hallmarks ( Clement, 2007 ; Colgrave et al. , 2020 ). Fibrosis is defined by the pathological accumulation of ECM proteins, resulting in scarring and thickening of the affected tissue ( Neary et al. , 2015 ). Historically, fibrosis was not listed as a hallmark of an endometriotic lesion, but the role of tissue remodelling that leads to fibrosis in humans ( Sillem et al. , 2001 ; Gilabert-Estelles et al. , 2003 ; Osteen et al. , 2003 ; Gilabert-Estelles et al. , 2005 ; Zhou and Nothnick, 2005 ; Balkowiec et al. , 2018 ) and animal model lesions ( Bruner et al. , 1997 ; Sharpe-Timms et al. , 1998 ; Bruner-Tran et al. , 2002 ; Osteen et al. , 2004 ; Stilley et al. , 2010 ; Stilley and Sharpe-Timms, 2012 ) is well described in the literature. Thus, our recommendations are updated to reflect the progression of the field to include fibrosis ( Vigano et al. , 2018 ). Additionally, lesions from any emerging or future model should be evaluated by an experienced endometriosis researcher and/or a gynaecological pathologist with experience in endometriosis to ensure the resulting histology meets these criteria. Schematic of an endometriotic lesion. Simplified view of an endometriotic lesion sectioned for histological assessment. Lesions characteristically have endometrial epithelial-lined glands, endometrial stroma, hemosiderin-ladened macrophages, and fibrotic tissue. Of the 14 evaluated protocols, key variables were identified, and some were unifiable; however, some could not be unified without jeopardizing future innovation. Key model variables were animal strain, location of ectopic tissue placement, volume of donor tissue, method of induction, hormonal status of recipient, and donor uterine tissue ‘type’. Included are SOPs for pre- and post-operative care of the mice (see Supplementary File S1 : EPHect-EM-Homologous-SOP 1 and 10). Below, our recommendations and supporting rationale are described ( Table 1 ). Advantages and disadvantages of prototypical homologous mouse models of endometriosis. Contains myometrium Decidualized tissue dominated by stromal content Mutations used are related to both endometriosis and endometrial cancer Contains myometrium Robust, > 8 months Robust, > 1 year with others noting early resolution Spontaneous resolution can start within 3 weeks Animals die or need to be euthanized Deep endometriotic 1-3 per animal 3-6+ per animal 1-4 per animal 3-6+ per animal Overabundance of smooth muscle/myometrium compared to human lesions Presence of epithelial glands is highly variable Overabundance of smooth muscle/myometrium compared to human lesions Presence of epithelial glands is highly variable Mutations present uniformly in all endometrial epithelium Vaginal bleeding and malignant characteristics may confound results Surgical induction increases innate immune cell infiltration Ovariectomy alters immune dynamics Surgery to develop the model, but no IP injection needed to disperse tissue Surgery required to implant the slow-release pumps containing SP and/or CGRP Defined mutations needed for induction Possible + Substance P and/or CGRP Possible with specific CRE driver Severe (vaginal bleeding, death); breeding mice with multiple mutations inefficient Reduction in body weight after implantation of slow release pumps Based on genetic manipulation : Yes; this is an advantage of this model. : No; this is a disadvantage of this model. : Maybe; the model is not the best for this. : Unknown; not enough data available to make a claim. For homologous mouse models, inbred/syngeneic strains are needed to avoid tissue rejection where uterine tissue is transplanted from one mouse (donor) into another (recipient); therefore, the basis for most mouse models include the use of C57BL/6 ( Efstathiou et al. , 2005 ; Hirata et al. , 2005 ; Becker et al. , 2008 ; Fainaru et al. , 2008 ; Matsuzaki et al. , 2008 ; Nowak et al. , 2008 ; Altan et al. , 2010 ; Jensen et al. , 2010 ; Körbel et al. , 2010 ; Lu et al. , 2010 ; Pelch et al. , 2010 ; Cheng et al. , 2011 ; Wilkosz et al. , 2011 ; Burns et al. , 2012 ; Pelch et al. , 2012 ; Wieser et al. , 2012 ; Tomio et al. , 2013 ; Cohen et al. , 2014 ; Greaves et al. , 2014b ; Kim et al. , 2014 ; Kumar et al. , 2014 ; Machado et al. , 2014 ; Pierzchalski et al. , 2014 ; Zhao et al. , 2014 ; Heard et al. , 2016 ; Peterse et al. , 2016 , 2018a ; Ruiz et al. , 2016 ; Dodds et al. , 2017 ; Ferrero et al. , 2017 ; Sanchez et al. , 2017 ; Burns et al. , 2018 ; Jones et al. , 2018 ; Yuan et al. , 2018 ; Chadchan et al. , 2019 ; Forster et al. , 2019 ; Horne et al. , 2019 ; Peyneau et al. , 2019 ; Sekiguchi et al. , 2019 ; Somigliana et al. , 1999 ; Tal et al. , 2019 ; Alali et al. , 2020 ; Fattori et al. , 2020 , 2024 ; Hattori et al. , 2020 ; Hayashi et al. , 2020 ; Li et al. , 2020 ; Mishra et al. , 2020 ; Symons et al. , 2020 ; Chen et al. , 2021 ; Ono et al. , 2021 ; Santorelli et al. , 2021 ; Sharma et al. , 2021 ; Dabrosin et al. , 2002 ; Zaninelli et al. , 2023 ) and BALB/c ( Bilotas et al. , 2010 ; Chang et al. , 2020 ; Cohen et al. , 2015 ; Dodds et al. , 2017 ; Mattos et al. , 2019 , Peyneau et al. , 2019 ; Ricci et al. , 2011 ; Ruiz et al. , 2016 ; Sanchez et al. , 2017 ; Uegaki et al. , 2015 ; Woo et al. , 2020 ; Yan et al. , 2019a ) mice. C57BL/6 mice are more commonly used in endometriosis research as most genetically modified strains are bred on the C57BL/6 background. Other inbred strains used for endometriosis studies, include FVB/N ( Dabrosin et al. , 2002 ; Dorning et al. , 2021 ; Jensen et al. , 2010 ), 129S6/SvEv ( Becker et al. , 2011 ), and CBA ( Peyneau et al. , 2019 ). The use of CD-1, ICR, and ddY ( Kusakabe et al. , 2010 ) mice have been described as well; however, these mice are outbred strains that lead to rejection of the uterine tissue, few lesions, and rapid lesion regression ( Lee et al. , 2009 ; Umezawa et al. , 2009 ; Kulak et al. , 2011 ; Silveira et al. , 2013 ; Liao et al. , 2014 ; Naqvi et al. , 2014 ; Li et al. , 2016 ; Wilson et al. , 2020 ), unless auto transplantation is performed ( Sharpe-Timms, 2002 ). Of note, the two most common strains of mice used, C57BL/6 and BALB/c, differ in their immunophenotype with C57BL/6 mice being Th1 dominant and BALB/c mice being Th2 dominant ( Watanabe et al. , 2004 ). A study by Dodds et al. (2017) compared these mouse strains and demonstrated that BALB/c mice had a greater proportion of cystic lesions, while the C57BL/6 mice had a greater variety of lesion types (i.e. cystic, dense, and necrotic). Endometriosis is both immune and hormonally regulated; therefore, immunological differences in the strain of mouse used could lead to diverse conclusions/findings unless these differences are accounted for when designing experiments or interpreting results. Behavioural studies related to animal activity and wellbeing can also differ between strains. For an in-depth discussion of strain-related behavioural responses, see the companion paper by Dodds et al. (2025 ). In those with endometriosis, lesions are most frequently identified within the peritoneal cavity, with (rare) extra-pelvic lesions found in the skin, pleural cavity, nasal cavity, and some other distant locations ( Sasmono et al. , 2003 ; Bagan et al. , 2008 ). Endometriosis is primarily a disease in women and those assigned female at birth; thus, it is modelled in female rodents. Historically, uterine tissue has been grafted both intraperitoneally ( Vernon and Wilson, 1985 ; Sharpe et al. , 1990 ; Sharpe et al. , 1991 ; Pelch et al. , 2010 ; Wilkosz et al. , 2011 ; Burns et al. , 2012 ; Pelch et al. , 2012 ; Greaves et al. , 2014b ; Mattos et al. , 2019 ; Tal et al. , 2019 ; Chang et al. , 2020 ; Hattori et al. , 2020 ; Li et al. , 2020 ; Symons et al. , 2020 ; Yoshino et al. , 2020 ; Chen et al. , 2021 ; Santorelli et al. , 2021 ), subcutaneously ( Cheng et al. , 2011 ; Wang et al. , 2013 ; Wang et al. , 2014 ; Ferrero et al. , 2017 ), and even on the hind leg in studies to evaluate pain mechanisms ( Bove, 2016 ). Subcutaneous engraftment can enable easier measurement of lesion development using callipers for external measurement and facilitating recovery of the lesions at the end of an experiment. This can be particularly beneficial for studies investigating the effects of therapeutics or interventions on lesion size/growth. However, the subcutaneous method does not recapitulate authentic interactions between endometrial and peritoneal tissues and the peritoneal immune microenvironment. Therefore, lesions are most commonly established in the peritoneal cavity of the rodent to allow the uterine tissue to be exposed to the peritoneal niche and sites of attachment most similar to human endometriosis. Placement into the peritoneal cavity, while ideal for most models, can limit measurement (i.e. growth or regression) of lesions depending on location and the depth at which the lesion attaches. Fluorescence and bioluminescence imaging can, however, better enable these endpoints ( Burns et al. , 2021 ; Dorning et al. , 2021 ). Peritoneal cavity placement is the most physiologically relevant location for the establishment of endometriotic lesions in mouse models, and the working group was unanimous in their decision to recommend that uterine tissue be placed intraperitoneally (see Supplementary File S1 : EPHect-EM-Homologous SOP 5 and 6). Among the various models, the most commonly used amount of tissue was a donor to recipient ratio of 1:1, which equates to ∼40 mg of non-decidualized uterine tissue containing myometrium ( Dodds et al. , 2017 ). Studies that have evaluated different volumes of endometrial tissue have identified that a 1:1 ratio produces the most lesions in the peritoneal cavity ( Altan et al. , 2010 ; Burns et al. , 2012 ; Dodds et al. , 2017 ; Forster et al. , 2019 ; Horne et al. , 2019 ; Kim et al. , 2020 ; Woo et al. , 2020 ; Dorning et al. , 2021 ). Studies using a 1:2 ratio of one donor to two recipient mice often report fewer lesions ( Somigliana et al. , 1999 ; Hirata et al. , 2005 ; Yoshino et al. , 2006 ; Bacci et al. , 2009 ; Chen et al. , 2009 ; Jensen et al. , 2010 ; Pittaluga et al. , 2010 ; Itoh et al. , 2011 ; Takai et al. , 2013 ; Tomio et al. , 2013 ; Uegaki et al. , 2015 ; Sanchez et al. , 2017 ; Yan et al. , 2019a ; Fattori et al. , 2020 ; Ono et al. , 2021 ; Fattori et al. , 2024 ; Zaninelli et al. , 2023 ). As the former ratio generates more lesions, for studies focusing on lesion burden as a key endpoint, we recommend using one donor uterus to one recipient mouse. Standardization of tissue volume will improve comparisons across laboratories and treatment outcomes. We also acknowledge that other factors are likely to influence the number of lesions formed (e.g. presence or absence of prior immune challenge in the peritoneal cavity). The most common method of introducing uterine tissue into the peritoneal cavity of mice was via intraperitoneal injection. Initially, in early rodent models of endometriosis, uterine tissue was sutured onto the intestinal mesentery or to the peritoneal wall ( Cummings and Metcalf, 1995 ; Efstathiou et al. , 2005 ; Becker et al. , 2008 ; Fainaru et al. , 2008 ; Matsuzaki et al. , 2008 ; Lee et al. , 2009 ; Umezawa et al. , 2009 ). Importantly, Bain et al. (2020) indicated that peritoneal surgery significantly impacts the immune cell profile of the cavity, which may have an impact on lesion attachment and other long-term outcomes. In fact, surgery, especially open abdominal surgery (e.g. laparotomy), can induce stress and impair cell-mediated immunity, accelerating the development of endometriosis ( Liu et al. , 2016 ; Long et al. , 2016 ). More recent work has moved away from surgical methods to allow the uterine tissue to seed itself and form unassisted attachments to sites in the peritoneal cavity ( Dabrosin et al. , 2002 ; Burns et al. , 2012 ; Li et al. , 2016 ; Jones et al. , 2018 ; Alali et al. , 2020 ; Fattori et al. , 2020 ; Morris et al. , 2021 ). However, this trend may hinder the study of endometriosis lesion interactions attached to specific abdominal organs. The working group agreed that induction via the intraperitoneal injection method is preferable, and two main methods are used. The first method uses an 18G needle and/or trocar to inject minced uterine tissue into the peritoneal cavity (see Supplementary File S1 : EPHect-EM-Homologous SOP 3 and 5) at the midline above the bladder, using 1–2 punctures. This method does not require surgery for the recipient mouse. Mice injected via this method typically have ∼2–3 lesions found near the bladder, adipose tissue, parietal peritoneum, uterus, and intestines, ranging from cystic to dark dense lesions ( Hsu et al. , 2010 ; Greaves et al. , 2014b ; Yuan et al. , 2018 ; Fattori et al. , 2020 ) and do contain the traditional features of endometriotic lesions ( Fattori et al. , 2020 ). Alternatively, minced uterine endometrial tissue is injected and dispersed through a small dorsolateral incision (5 mm) that is closed with a clip with no suture to the peritoneal cavity ( Burns et al. , 2012 , 2018 ; Jones et al. , 2018 ; Morris et al. , 2021 ) (see Supplementary File S1 : EPHect-EM-Homologous SOP 3 and 6). In this model, neutrophils and macrophages increase 3–4-fold with the initiation of disease compared to sham-operated animals at 24, 48, and 72 h ( Burns et al. , 2018 ). This model develops ∼3–6+ lesions per mouse that begin as white tissue and are red/haemorrhagic by 48–72 h. By 2 weeks, lesions are cystic and exhibit organized glands, stroma, and haemosiderin macrophages, and are fibrotic ( Burns et al. , 2018 ). These lesions are found attached to ovarian adipose tissue, intestines, cul-de-sac, peritoneal wall, stomach, diaphragm, uterus, and bladder and are maintained for up to a year (experimental testing was stopped at 1 year) in this mouse model ( Jones et al. , 2018 ). Alternatively, one research group described endometriosis induction via laparoscopy ( Peterse et al. , 2016 , 2018b , 2024 ). The advantage of laparoscopically guided inoculation is that uterine tissue can be placed at locations where lesions are most often removed in women, namely the peritoneal wall and cul-de-sac. Using this model, up to 60% of the implanted uterine tissue pieces could be retrieved after 1 week ( Peterse et al. , 2016 , 2018a , 2024 ). The main disadvantage of this model is that it requires specialized equipment and a highly skilled researcher to perform the laparoscopy ( Corona et al. , 2011 ). Endometriosis is a hormonally responsive disease in women, where endometriotic lesions are exposed to cyclical fluctuations of reproductive hormones throughout the menstrual cycle. A large number of mouse studies use exogenous hormones (oestrogen alone, oestrogen followed by progesterone, or pregnant mare serum gonadotropin (PMSG) to synchronize the donor uterine tissue prior to being used to initiate endometriosis ( Dabrosin et al. , 2002 ; Hirata et al. , 2005 ; Yoshino et al. , 2006 ; Bacci et al. , 2009 ; Altan et al. , 2010 ; Jensen et al. , 2010 ; Pittaluga et al. , 2010 ; Itoh et al. , 2011 ; Burns et al. , 2012 ; Wieser et al. , 2012 ; Takai et al. , 2013 ; Tomio et al. , 2013 ; Greaves et al. , 2014b ; Pierzchalski et al. , 2014 ; Kim et al. , 2014 , 2020 ; Liao et al. , 2014 ; Ruiz et al. , 2016 ; Sanchez et al. , 2017 ; Burns et al. , 2018 ; Jones et al. , 2018 ; Peterse et al. , 2018a ; Yuan et al. , 2018 ; Forster et al. , 2019 ; Horne et al. , 2019 ; Sekiguchi et al. , 2019 ; Somigliana et al. , 1999 ; Yan et al. , 2019a ; Alali et al. , 2020 ; Fattori et al. , 2020 ; Li et al. , 2016 , 2020 ; Dorning et al. , 2021 ; Ono et al. , 2021 ; Fattori et al. , 2024 ; Zaninelli et al. , 2023 ) (see Supplementary File S1 : EPHect-EM-Homologous SOP 2, 3, and 4). Other studies have taken donor uterine tissue in oestrus ( Nowak et al. , 2008 ; Wang et al. , 2013 ; Dodds et al. , 2017 ); however, this method requires many donor animals on hand to ensure oestrus on the day of surgery. Due to these differences in study design, our recommendation is that investigators be consistent in their studies to use hormone-supplemented donors or to use cycle-staged donors. For the recipient, we recommend that experimental endometriosis is established in mice with intact ovaries to mimic the cyclical hormonal fluctuations that naturally occur. Early mouse models of endometriosis often utilized ovariectomy (see Supplementary File S1 : EPHect-EM-Homologous SOP 9) with supplementation of oestrogen to reduce hormonal variation. Studies have found that exogenous oestrogen increases lesion size but not lesion number ( Burns et al. , 2012 ; Burns et al. , 2018 ; Jones et al. , 2018 ; Morris et al. , 2021 ). Nevertheless, using intact mice has become increasingly common to more closely mimic human disease and avoid surgery. For studies evaluating hormones or hormonal regulation, ovariectomy followed by hormonal supplementation may be required as part of the experimental design, although natural cycling is required when testing potential therapeutics to determine their impact on cyclicity. When using intact animals, Dodds et al. (2017) found a slight difference in the number of lesions that attached when tissue was injected during oestrus (relatively lower oestradiol levels) versus the proestrus phase (relatively higher oestradiol levels) of the cycle; metestrus and dioestrus were not examined. In contrast, other investigators have not observed any correlation between lesion number and cycle stage at tissue injection (( Dorning et al. , 2021 ) and unpublished data). While the cycle stage at induction may have minor impacts on tissue attachment, the cycle stage at necropsy is more important to control as hormone status (see Supplementary File S1 : EPHect-EM-Homologous SOP 8) impacts expression of genes and proteins in endometriosis lesions (i.e. progesterone receptor (PGR), matrix metalloproteinases (MMPs), lactoferrin (LTF), etc.) ( Burns et al. , 2012 ). With the advent of static isolated cages, mice are no longer exposed to male pheromones, and consequently do not cycle as regularly; therefore, we recommend that male bedding be placed in the female cages every other day to every third day to expose the female mice to pheromones secreted by the male mice ( Whitten et al. , 1968 ; McCarthy et al. , 2018 ). To allow for innovation of experimental endometriosis models in the future, we cannot recommend standardization of the donor uterine tissue ‘type’. However, it is appropriate to discuss different types of tissue currently being used to induce experimental endometriosis. The aetiology of endometriosis remains unknown, but most models begin with uterine tissue with/without myometrium, with the aim of recapitulating retrograde menstruation or at least to establish ectopic endometrial-like tissue engraftment. In support of this approach, other tissue types (e.g. lung, mammary gland, and bladder) minced and injected into the peritoneal cavity do not attach or form endometriosis-like lesions; however, other investigators show that adipose tissue does attach and can attract blood vessels and nerve fibres but does not develop the typical endometriotic features of glands or stroma ( Castro et al. , 2021 ; Peterse et al. , 2024 ). Importantly, menstrual effluent from women consists of epithelial cells, decidualized stromal cells, and multiple immune cell types ( Salamonsen and Lathbury, 2000 ; Salamonsen, 2021 ). To represent these cell types likely involved in endometriosis development, three donor uterine tissue ‘types’ have emerged for use in the mouse model, all with their own pros and cons. The use of full-thickness uterine tissue produces endometriosis-like lesions that are often cystic. These lesions have epithelial-lined glands, stroma, and hemosiderin-laden macrophages and exhibit fibrosis ( Lu et al. , 2010 ; Peterse et al. , 2016 ; Ferrero et al. , 2017 ; Yan et al. , 2019a ; Fattori et al. , 2020 ). Importantly, lesions remain viable and robust in the mouse for long-term studies ( Dorning et al. , 2021 ). An advantage to this model is that specific skills are not required to prepare the donor tissue or to inject minced uterine tissue, making the full-thickness model an attractive model for ease and higher throughput of animals. Lesions derived from full-thickness uterine tissue frequently contain the key hallmarks of lesions, are larger, and exhibit a greater bioluminescent signal in a model implemented for non-invasive monitoring of lesions. Compared to a model utilizing decidualized ‘menses-like’ endometrium, ∼35% more lesions were evident 6 weeks after model induction ( Dorning et al. , 2021 ); thus, the full-thickness model exhibits significantly less spontaneous resolution of lesions. Functionally, the lesions are hormone- and treatment-responsive in growth and gene expression changes. Mice exhibit an increase in both spontaneous and evoked pain-like behaviours compared to sham animals ( Lu et al. , 2010 ; Fattori et al. , 2020 ). A weakness of the model is that the injection of full-thickness uterine tissues results in the incorporation of myometrial tissue into lesions. While endometriosis lesions in women often show staining for smooth muscle actin ( Anaf et al. , 2000 ; Barcena de Arellano et al. , 2011 ) and the presence of cells with a myofibroblast phenotype, it is unlikely that the tissue type injected and the amount of muscle tissue present in this model fully mimic human disease. As with all models, the amount of tissue injected into the peritoneal cavity does not correspond to the number of lesions formed. The lesions are variably located, of variable sizes, and can be difficult to locate and quantify if fluorescence/bioluminescence is not used. The preparation of the uterine tissue for implantation in this model is straightforward, easy, and fast, which allows for inducing large numbers of animals for generating dose–response curves to test the efficacy of potential drugs. The lesions produced in this model have been induced by either injecting minced uterine tissue from donor animals into the peritoneal cavity of genetically identical recipient mice using an 18G needle (see Supplementary File S1 : EPHect-EM-Homologous SOP 5) ( Fattori et al. , 2020 ) or by performing laparotomy and suturing uterine tissue onto the bowel mesentery, similar to what has been described in Supplementary File S1 : EPHect-EM-Homologous SOP 12 ( Bilotas et al. , 2010 ; Pelch et al. , 2010 ), although the latter has become obsolete in mice in recent years. Endometriosis mouse models induced with this type of tissue have been used to study early development of endometriosis lesions, angiogenesis, endometriosis-associated pain, and endometriosis-associated infertility, and to identify new therapeutic treatment options or to examine the efficacy of re-purposed FDA-approved drugs ( Lin et al. , 2006 ; Bilotas et al. , 2010 ; Lu et al. , 2010 ; Ruiz et al. , 2016 ; Fattori et al. , 2020 ; Elsherbini et al. , 2022 ; Maddern et al. , 2022 ) for endometriosis. In addition, investigators have used this model to optimize a laparoscopic procedure for endometriosis induction as well as to investigate immune changes in the peritoneum in response to lesion development ( Peterse et al. , 2016 ; He et al. , 2022 ). Other investigators have used this model to study neuroimmune communication in endometriosis. They showed that in mice, nociceptor ablation reduced pain, monocyte recruitment, and lesion size ( Fattori et al. , 2024 ). In general, this model is very versatile and can be used to study pathogenesis as well as the pathophysiology of endometriosis. Similar to the full-thickness model, obtaining endometrial tissue by peeling away the myometrium in layers from the outside of the uterus (as opposed to scraping the endometrium away from the myometrium) or using sharp dissection to isolate the endometrial layer ( Dodds et al. , 2017 ; Dorning et al. , 2021 ) produces lesions that are often cystic, dense, and/or haemorrhagic at different stages post-inoculation ( Burns et al. , 2018 ). Lesions are typically found attached to uterine arteries, cul-de-sac, fat pads or peritoneal wall, or near the spleen/stomach area and intestine; however, this can be somewhat dependent on the angle of tissue injection. Lesions are typically not found on the liver, kidney, or ovary. These lesions have epithelial-lined glands, stroma, fibrosis, and hemosiderin, but have less myometrium than the full-thickness model ( Burns et al. , 2012 ; Dodds et al. , 2017 ; Burns et al. , 2018 ; Jones et al. , 2018 ; Dorning et al. , 2021 ; Morris et al. , 2021 ). Lesions derived from the endometrium model survive a minimum of 2 months. Lesion number, size, and appearance are variable, based on the method used to dissect the endometrium before implantation. Functionally, lesion growth and gene expression changes are hormonally responsive ( Burns et al. , 2012 , 2018 ; Jones et al. , 2018 ). Animals in this model exhibit pain-like symptoms, with a reduction in abdominal reaction threshold in response to von Frey filaments (mechanical hyperalgesia), but no differences in paw withdrawal was observed compared to sham-operated animals ( Dorning et al. , 2021 ). As with all injection models, one weakness is that the amount of tissue inoculated into the peritoneal cavity does not necessarily correspond to the number of lesions formed. In addition, the lesions are variably located, are of variable sizes, and can be difficult to locate if fluorescence/bioluminescence is not used. The use of endometrial tissue for the development of endometriosis lesions recapitulates the appearance of endometriotic lesions with the presence of epithelial cells, glands, and stroma. The lesions from this model are induced by injecting homogenized endometrial tissue by needle (18G) or surgically injecting minced endometrium into the peritoneal cavity of the mouse ( Burns et al. , 2012 ; Dodds et al. , 2017 ; Burns et al. , 2018 ; Jones et al. , 2018 ; Dorning et al. , 2021 ; Morris et al. , 2021 ) (see Supplementary File S1 : EPHect-EM-Homologous SOP 3, 5, or 6). While this model requires trained individuals to prepare the donor endometrium, this version of the model offers several benefits; lesions are found in the peritoneal cavity, are larger or more numerous than some models, survive in the peritoneal cavity long-term, develop into mature lesions, and are varied in size and location. Lesions found in the peritoneum range from dense to cystic lesions that are responsive to hormones, therapeutics, and/or endocrine-disrupting compounds ( Burns et al. , 2012 ; Dodds et al. , 2017 ; Burns et al. , 2018 ; Jones et al. , 2018 ; Dorning et al. , 2021 ; Morris et al. , 2021 ). This model can and has been used to study early implantation of lesions, immune-mediated changes in response to lesion development, lesion survival, and changes in pain sensory behaviour. In humans, the functional layer of the endometrium, including endometrial epithelial and stromal cells, undergoes spontaneous decidualization and is shed during menstruation. In contrast, rodents have an oestrous cycle, the endometrium does not spontaneously decidualize ( Ng et al. , 2020 ), and the endometrium is not shed but is instead remodelled during the cycle. To recapitulate menstruation, donor mice undergo a truncated human hormonal schedule, artificial decidualization is induced by injecting sesame or corn oil into a uterine horn, and shedding is initiated by progesterone withdrawal ( Finn and Pope, 1984 ; Cousins et al. , 2014 ; Greaves et al. , 2014b ; Peterse et al. , 2018b ; Horne et al. , 2019 ; Forster et al. , 2019 ; Kim et al. , 2020 ) (see Supplementary File S1 : EPHect-EM-Homologous-SOP 4). The decidualized donor tissue is used to establish lesions in recipient mice with the aim of mimicking retrograde menstruation. In contrast to human decidualized tissue, which does maintain glandular structures, a drawback to the decidualized rodent endometrium is that it contains mostly decidualized stromal cells, with few epithelial glands ( Peterse et al. , 2018b , 2024 ). Moreover, variation has been observed in the degree of decidualization and the amount of decidualized endometrium in donor mice. In mice, during decidualization, there are often areas of the endometrium that lack decidualization; thus, these non-decidualized areas still contain the normal endometrial glandular structures ( Finn and Pope, 1984 ; Brasted et al. , 2003 ; Xu et al. , 2007 ; Rudolph et al. , 2012 ; Peterse et al. , 2018b , 2024 ). In recipient mice, endometriosis lesions in this model are easy to identify as they are often red at 2–3 weeks post induction. The lesions are found attached to the peritoneal wall, fat, bladder and, occasionally, the intestines. Not all the injected tissue attaches, and floating material can be identified during laparotomy for lesion recovery ( Greaves et al. , 2014b ; Dorning et al. , 2021 ). Some lesions in this model exhibit spontaneous regression from 3 weeks post tissue inoculation and do not contain as many epithelial gland structures as models that use naïve endometrium or full-thickness uterine tissue ( Dorning et al. , 2021 ). Other investigators indicated that endometriosis lesions in this model rarely contained epithelial glands but do often contain hemosiderin-laden macrophages ( Peterse et al. , 2018a , 2024 ). Levels of fibrosis are comparable to other models ( Dorning et al. , 2021 ), indicating this is a consistent feature of lesions recovered from other experimental endometriosis models. The use of decidualized stroma for the development of endometriosis lesions in mice represents the development of endometriosis lesions from endometrium, including decidualized stroma, in humans. The differentiation process of decidualization does not happen spontaneously in mice; thus, decidualized tissue is acquired through hormonal and mechanical processes. The use of ‘menses-like’ endometrium in the peritoneal cavity to develop endometriosis lesions can be done with or without hormone replacement in the recipient and allows for early implantation studies and studies investigating the contribution of the immune system ( Forster et al. , 2019 ; Cousins et al. , 2020 ; Hogg et al. , 2020 , 2021 ; Dorning et al. , 2021 ; Greaves et al. , 2014a , b , 2015 , 2017b ). Using this model, mice exhibit pain-like behaviours and molecular alterations in the peripheral and central nervous system 2–3 weeks post lesion induction that mimic human disease and associated maladaptation in the nervous system ( Greaves et al. , 2017b ). Lesions are infiltrated by abundant macrophages that drive lesion pathogenesis ( Hogg et al. , 2020 ) and are associated with the generation of pain-like behaviours ( Forster et al. , 2019 ). Macrophage infiltration as well as neuroangiogenesis are regulated by oestrogen signalling in this model ( Greaves et al. , 2014a , 2015 ). Moreover, Greaves et al. have used this model to identify a protective population of monocyte-derived peritoneal macrophages that, when present in abundance, result in fewer and smaller lesions ( Hogg et al. , 2021 ). This supports the theory that women with endometriosis exhibit immune dysfunction and that aberrations in macrophage function are linked with lesion persistence. This preclinical model has therefore also been used to test the repurposing potential of drugs that target metabolic processes as a non-hormonal treatment for endometriosis ( Horne et al. , 2019 ). From the evaluated protocols and publications on the rat model of endometriosis, far fewer differences were found among the protocols than were identified in publications using mice. Key variables for the rat model were strain, location of ectopic tissue placement, volume of tissue, and hormonal status. Methods of induction and tissue type were mostly similar in all protocols. Below, we describe our recommendations and their rationale ( Table 2 ). Advantages and disadvantages in prototypical homologous rat models of endometriosis. Contains myometrium Decidualized tissue dominated by stromal content When lesions are present + 2 months + 3 months when present When lesions are present 4 sutured per animal 1-4 per animal when present Overabundance of smooth muscle/myometrium compared to human lesions Epithelial glands are not common Epithelial glands are not common Spontaneous ongoing pain symptoms Uterine tissue is sutured into intestinal mesentery : Yes; this is an advantage of this model. : No; this is a disadvantage of this model. : Unknown; not enough data available to make a claim. Most studies in rat, to date, have used the outbred albino Sprague-Dawley strain, likely due to their docility, ease of handling, regular reproductive cyclicity, and the ability to develop transgenic models ( Sharpe-Timms, 2002 ). Far fewer studies have used the Wistar rat (also outbred), which tend to be slightly smaller in size. An autologous rat (Sprague-Dawley) surgical model was originally developed in 1985 ( Vernon and Wilson, 1985 ). Four full-thickness uterine squares from the distal third of one uterine horn were auto-transplanted onto the arterial cascades of the small intestine (see Supplementary File S1 : EPHect-EM-Homologous SOP 11, 12, and 15). The implant size increased over time, reaching maximal growth at 60 days post transplantation ( Vernon and Wilson, 1985 ). The endometriosis implants developed into cyst-like structures and, depending on the stage of the oestrous cycle or hormonal treatments, could be fluid-filled. Initial studies compared the growth of the implants to sham-controls and to the injection of uterine lavage or endometrial scrapings into the peritoneal cavity but found that the only viable endometriosis implants were those from the surgical implantation procedure ( Vernon and Wilson, 1985 ). With this knowledge, more recent studies compared endometrial tissue implants to fat implants or suture ties ( Sharpe-Timms, 2002 ). Other variations of rat models for endometriosis include subcutaneous placement of endometrial tissue fragments. An initial study placed endometrial tissue sections subcutaneously in Sprague-Dawley rats to examine implant regression after ovariectomy and during pregnancy ( Vernon and Wilson, 1985 ). A second study implanted tissues onto the muscle of the inguinal region in Wistar rats to demonstrate the effects of drugs on reducing implant size ( Pereira et al. , 2015 ). In a third study, subcutaneous endometrial tissue fragments in Wistar rats were used to evaluate the effect of oestrogenic properties of avocado seed extract on macroscopic implant regression ( Minko Essono et al. , 2020 ). However, to account for consistency in lesion size, distal proximity to the utero-ovarian vasculature, and development of the lesions produced, the placement of fragments on the arterial cascades is the recommended option. Studies using the auto-transplantation model modified the number of implants from four to six full-thickness uterine squares and initiated studies at 4 weeks post auto-transplantation ( Vernon and Wilson, 1985 ; Sharpe et al. , 1990 ; Sharpe et al. , 1991 ; Wright and Sharpe-Timms, 1995 ; Sharpe-Timms et al. , 1998 ; Cox et al. , 2001 ; Sharpe-Timms, 2002 ; Stilley et al. , 2009 ; Stilley and Sharpe-Timms, 2012 ; Birt et al. , 2013b ; Sharpe-Timms et al. , 2020 ). Whether using four or six implants, the lesions remain viable for approximately 1 year in duration if initially established in young adult animals. Despite the surgical nature of this model, studies of developing endometriotic implants in rat uterine auto-transplantation models aided the discovery of mechanisms of the genesis of endometriotic implants. For example, endometriotic implants at 36 h, 2 weeks, and 4 weeks post-surgery progressed from pale avascularized implants to red vascularized lesions, to well-established encapsulated cysts that express enzymes, playing a role in peritoneal remodelling and progression of the implants. Others have also observed angiogenesis as well as neurogenesis developing as early as 7 days after surgery ( Torres-Reveron et al. , 2018a ). As with the mouse model, the growth of endometrial tissue is oestrogen-dependent; thus, the recommended hormonal status of the rat is to be hormonally intact unless the research goal is to study hormones or hormonal regulation. Like humans, regression of uterine ectopic implants can be induced in rats by generating a hypo-oestrogenic state, either by ovariectomy, application of GnRH agonists or synthetic compounds ( Sharpe et al. , 1990 ; Sharpe et al. , 1991 ; Wright and Sharpe-Timms, 1995 ; Sharpe-Timms et al. , 1998 ; Yavuz et al. , 2007 ; Yuan et al. , 2012 ). Proper cycling must be ascertained before endometriosis induction and should be monitored at various points during the experimental protocol if therapeutic interventions are being tested to account for impact on cyclicity (whether on alterations in length of cycle or proportion of time spent in each stage). Like mouse models, the oestrous cycle stage at necropsy must be determined if examining hormonally responsive changes (see Supplementary File S1 : EPHect-EM-Homologous SOP 14). If ovariectomized rats are used, supplementing with oestradiol is recommended for lesion growth. The rat model has been useful for studies of pain, fertility, stress, hormone modulation, ovarian function, immunomodulation, inflammation, gene expression and epigenetic regulation, protein synthesis and secretion, therapeutic modulation of the implants, and others ( Vernon and Wilson, 1985 ; Rajkumar et al. , 1990 ; Barragan et al. , 1992 ; Moon et al. , 1993 ; Sharpe and Vernon, 1993 ; Nothnick et al. , 1994 ; Sharpe-Timms et al. , 1998 ; Sharpe-Timms, 2002 ; Uchiide et al. , 2002 ; Cason et al. , 2003 ; Berkley et al. , 2004 ; Rojas-Cartagena et al. , 2005 ; Berkley et al. , 2007 ; Flores et al. , 2007 ; Zhang et al. , 2008a ; McAllister et al. , 2009 ; Stilley et al. , 2009 , 2010 ; Cuevas et al. , 2012 ; Birt et al. , 2013a , b ; Appleyard et al. , 2015 , 2020 , 2021 ; Hernandez et al. , 2015 , 2017 ; Torres-Reveron et al. , 2016 , 2018a,b, 2020 ; Cuevas et al. , 2018 ; Lopez et al. , 2020 ; Seguinot-Tarafa et al. , 2020 ). The rat model of endometriosis mirrors many end-points of the disease in patients, including ectopic lesions, impaired natural killer cell activity, protein synthesis, and secretion by the ectopic lesions ( Sharpe and Vernon, 1993 ; Sharpe et al. , 1993 ), increased inflammatory mediators (in peritoneal fluid, serum, lesions, and endometrium), elevation of angiogenic molecules, upregulation of nerve growth factors and their receptors, damage to the intestinal tract, increased vaginal hyperalgesia, and changes in the central pathways of pain modulation ( Vernon and Wilson, 1985 ; Moon et al. , 1993 ; Cason et al. , 2003 ; Berkley et al. , 2004 , 2007 ; Grummer, 2006 ; Zhang et al. , 2008b ; McAllister et al. , 2009 ; Cuevas et al. , 2018 ). Endometriotic implants in rats are histologically alike, respond to steroids, and activate similar inflammatory mechanisms to those that characterize human disease, with a comparable global gene expression profile ( Nothnick et al. , 1994 ; Sharpe-Timms, 2002 ; Uchiide et al. , 2002 ; Rojas-Cartagena et al. , 2005 ; Flores et al. , 2007 ; Seguinot-Tarafa et al. , 2020 ). Compared to the mouse, the rat appears better suited to investigations on fertility and those using complex behavioural studies. The rat model is well suited to evaluations incorporating aspects of gastrointestinal physiology and cross-organ sensitization due to the lesions being sutured on to the intestinal mesentery ( Torres-Reveron et al. , 2016 ; Lopez et al. , 2020 ; Torres-Reveron et al. , 2018a , 2020 ). The contribution of stress to the development and progression of endometriosis is also observed in the rat model ( Cuevas et al. , 2012 ; Appleyard et al. , 2015 , 2020 ; Hernandez et al. , 2017 ), as well as the positive contribution of environmental enrichment and exercise ( Hernandez et al. , 2015 ; Torres-Reveron et al. , 2018b ; Appleyard et al. , 2021 ). An important attribute to using a rat model is that many pharmacokinetic studies are initially done in rats. For drug discovery work, traditional adsorption, distribution, metabolism and excretion, work is often done in rats to examine toxicity and pharmacokinetics of the compound. Somatic cancer driver mutations are observed across all subtypes of endometriosis ( Anglesio et al. , 2017 ; Chui et al. , 2017 ; Guo, 2018a ; Suda et al. , 2018 , 2019 ; Bulun et al. , 2019 ; Lac et al. , 2019b ; Praetorius et al. , 2022 ; Orr et al. , 2023 ). While significant differences between subtypes have not been described, smaller cohorts suggest that ovarian endometriomas have greater mutation burdens and increased levels of mutational heterogeneity ( Praetorius et al. , 2022 ). Mutational analysis shows that recurrent cancer driver mutations are found exclusively in the epithelial compartment of endometriotic lesions and normal-appearing endometrial tissue ( Lac et al. , 2019a ; Moore et al. , 2020 ; Yamaguchi et al. , 2022 ). In support, somatic non-driver mutations observed in the stromal compartment of lesions are less frequent and do not appear to be recurrent ( Lac and Huntsman, 2018 ; Noe et al. , 2018 ; Suda et al. , 2019 ). The invasive phenotype of endometriosis shares aspects of tumour metastasis and invasiveness, including migration, anchorage-independent growth, recruitment of vasculatures, and survival at distal sites ( Giudice and Kao, 2004 ; Cheng et al. , 2011 ; Wilson et al. , 2020 ). Cancer driver mutations are thought to have invasive cell properties and to promote lesion establishment and growth. For example, patients with more severe stages of endometriosis are more likely to have lesions that contain cancer driver mutations ( Praetorius et al. , 2022 ; Orr et al. , 2023 ). Importantly, somatic mutations are not exclusively a marker of malignant potential. Mutant mouse models carrying cancer driver mutations support a role for these mutations in the establishment and spread of invasive endometrial-like tissue ( Dinulescu et al. , 2005 ; Cheng et al. , 2011 ; Wilson et al. , 2020 ). Given the natural tendency of cancer driver mutations to generate invasive cell properties, which could have selective advantages for endometrial tissue in the peritoneal cavity, lesion establishment may not be as dependent on endometrial tissue transplantation ( Wilson et al. , 2020 ) or the use of exogenous hormones ( Cheng et al. , 2011 ) in these mouse models. Cancer driver mutations may promote additional aspects of disease pathogenesis, such as changes to the lesion microenvironment and hormone responsiveness. The existence of genetically engineered mouse or rat strains to study cancer-related processes in vivo offers the opportunity to study the effects of somatic mutations and other genetically traceable alterations on pathogenic mechanisms in endometriosis rodent models ( Dinulescu et al. , 2005 ). Additional improvements to existing models or the establishment of new models may allow for time-based control of mutation induction or the ability to study cancer driver mutations specifically in ectopic lesions, avoiding the unnecessary phenotypic effects of mutations in eutopic endometrium. For example, early evidence for the role of cancer-associated mutations in endometriosis came from the inadvertent induction of endometriosis-like lesions in an ovarian cancer mouse model, by injecting a cre-expressing virus into the LSL-KRAS(G12D) transgenic mouse strain ( Dinulescu et al. , 2005 ). Transplant models offer the ability to study cancer driver mutations in ectopic lesions, avoiding the confounding phenotypic effects of mutations in eutopic endometrium. On the other hand, somatic cancer driver mutations are not only observed in endometriosis lesions but have also been found in histopathologically normal-appearing endometrium of women with and without endometriosis. It is currently unknown whether cancer driver mutations in eutopic endometrium are causative of endometriosis development, thus, co-modelling of ectopic and eutopic genetically modified endometrium may be desirable. One subtype of endometriosis that often causes severe pain in those with endometriosis and poses a challenge in clinical management is deep endometriosis ( Tosti et al. , 2015 ). While a baboon deep endometriosis model was established a decade ago, the model requires facilities and expertise for the use of non-human primates ( Donnez et al. , 2013 ). This model is cost-prohibitive and may have similar experimental limitations as humans; hence, the use of this model is not feasible for most laboratories. Capitalizing on the finding that neuropeptides, such as substance P and calcitonin gene-related peptide (CGRP), secreted by sensory nerves surrounding the endometriotic lesions actively facilitate lesion progression and fibrogenesis ( Liu et al. , 2019 ; Yan et al. , 2019b ), a deep endometriosis mouse model can be established by chronic infusion of substance P and/or CGRP in a conventional endometriosis mouse model. This model differs from the conventional rodent models of endometriosis by fulfilment of the four basic requirements for deep endometriotic lesions: (i) the resultant lesions consist of endometrial stromal and epithelial cells, (ii) lesions are encapsulated in surrounding tissues or organs (e.g. bowel wall, rectovaginal septum), (iii) lesions display signs consistent with smooth muscle metaplasia, and (iv) lesions exhibit extensive fibrosis ( Yan et al. , 2019a ). As such, this model offers a viable and much more economical option than the baboon deep endometriosis model. In addition, this model is based on known biological mechanisms underlying the deep endometriosis formation. It also provides an experimental explanation for a key difference between deep endometriosis and ovarian endometrioma: deep endometriosis lesions in humans are always located, without exception, in proximity to various nerve plexuses in the pelvic cavity ( Foti et al. , 2018 ). In recent years, rat injection models have begun to be developed in a similar manner to the mouse injection model. Persoons et al. (2020) induced endometriosis in adult Sprague-Dawley rats, adapting a protocol based on the menstruating mouse model. In brief, menstrual endometrial tissue separated from the myometrium obtained from the decidualized uterine horn of donor rats was finely minced and injected intraperitoneally into recipients with synchronized oestrus cycles using a 16G catheter ( Persoons et al. , 2020 ). Here, a 1:1 ratio of donor to recipient tissue was used to induce the disease. Endometriosis-like lesions could be detected in 45% of the recipient animals after 12 weeks. Similar to the mouse model, unattached tissue can be observed in the abdominal cavity up to 12 weeks after seeding. The lesions varied in colour, abdominal location, and size. Notably, most of the lesions were located near connective tissue, intestines, or gonadal adipose tissue. Lesion size was considerably smaller than the suture model, and, in most cases, only one lesion could be retrieved. Interestingly, rats exhibit spontaneous postural changes that reflect ongoing pain behaviour, making it a useful model to study pain pathophysiology and potential treatments. Alternatively, recipient Sprague-Dawley rats were injected intraperitoneally with pooled minced donor tissue (equivalent to one uterus) along the midventral line using an 18G needle ( Cordaro et al. , 2021 ; D'Amico et al. , 2022a , b ; Genovese et al. , 2021 , 2022 ; Siracusa et al. , 2021a , b ). The donor rats were synchronized by administering 10 IU PMSG. The rats were anaesthetized to harvest the uterus, ovary, and oviducts, and the uterine horn was minced in phosphate-buffered saline (PBS) prior to injection. Lesions were allowed to develop for 14 days, producing an average of two per recipient ∼1–1.5 mm in diameter. Disadvantages to the rat injection model include limitations on identifying the location of the lesions, the growth of the lesions, and minimal development of multiple lesions. The use of inbred strains would be expected to improve the number and duration of lesion implants. To reduce the number of animals required to answer research objectives, we recommend that all studies include outcomes, lesion location, and lesion histology (see below) as well as several of the following endpoints. The inclusion of consistent outcomes will allow for harmonization across rodent endometriosis model studies and facilitate collaboration and hypothesis generation by cross-comparison. Bias should be minimized when making measurements. Rodents should be randomized to each experimental group, and the measuring investigator should be blinded to the experimental status of the animals as much as possible, with data analysis decisions (e.g. exclusion of outliers) being finalized before unblinding. Finally, appropriately powered sample sizes should be used and based on the variance, once known. The extent to which these procedures are employed must be made clear in both the laboratory notebook and the manuscript. Documentation of accurate lesion location is recommended. This documentation will allow investigators to capture lesion metrics for each animal. We provide a rodent body diagram that is helpful when marking lesion location in the peritoneal cavity ( Fig. 2 ). On these sheets, for each animal, other metrics can be documented. We recommend the following: total lesion weight (each lesion alone may be too small to weigh), volume of each lesion as determined using callipers to take height, width, and length (mm 3 ), lesion location and point of attachment, lesion colour, and presence of vascularization. Other important metrics include findings in the peritoneal cavity, such as the presence of adhesions, swollen lymph nodes, and/or the abnormal appearance of other organs. Diagram for use in necropsy to locate endometriotic lesions. The use of a diagram for each animal will allow for standardization of findings to indicate lesion location, size (weight and volume), colour, etc. The diagram can also be used to note other normal or abnormal findings (e.g. uterine weight, ovary weight, liver appearance). Intestinal and adipose tissues are not depicted. For any new model, the histology of lesions (e.g. H&E staining) must be carefully examined and verified in collaboration with an experienced endometriosis research scientist and/or pathologist with extensive experience in endometrial/endometriotic histology to determine the extent to which lesions resemble human endometriosis. Routine confirmation of lesion histology is advisable for all investigators but is especially important for new investigators in a laboratory, who may mistake other structures (e.g. lymph nodes or mammary tissue) for lesions. Fluorescence/bioluminescence imaging of lesions for localization in vivo allows for the longitudinal monitoring of the number and size of lesions as well as visualization of lesion location prior to necropsy, although this can be hampered by the length of the protocol and autofluorescence from gastrointestinal content ( Dorning et al. , 2021 ). Photographic documentation of lesions using white light visualization at necropsy allows for lesion colour, approximate size, and localization, if imaged while in the animal. Ex vivo imaging will also enhance visualization for future resources and/or comparisons among laboratories. Endometriosis is known to adversely affect fertility and pregnancy outcomes in women; thus, assessment of these endpoints is appropriate for some studies ( Macer and Taylor, 2012 ; Stilley et al. , 2012 ; Prescott et al. , 2016 ; Farland et al. , 2019 ; Breintoft et al. , 2021 ). Most fertility studies in rodents with endometriosis have been done in rat models of endometriosis ( Vernon and Wilson, 1985 ; Barragan et al. , 1992 ; Moon et al. , 1993 ; Sharpe-Timms et al. , 1998 ; Sharpe-Timms, 2002 ; Stilley et al. , 2009 ; Stilley et al. , 2010 ; Stilley and Sharpe-Timms, 2012 ; Birt et al. , 2013a , b ; Sharpe-Timms et al. , 2020 ; Kanellopoulos et al. , 2022 ). The homologous induction of endometriosis in rats has been used to characterize many defects that cause subfertility ( Vernon and Wilson, 1985 ; Barragan et al. , 1992 ; Moon et al. , 1993 ; Sharpe-Timms et al. , 1998 ; Sharpe-Timms, 2002 ; Stilley et al. , 2009 ; Stilley et al. , 2010 ; Stilley and Sharpe-Timms, 2012 ; Birt et al. , 2013a , b ; Sharpe-Timms et al. , 2020 ; Kanellopoulos et al. , 2022 ). In the earliest studies, endometrial tissue transplantation reduced both the number of pups at term (by 48%) and number of Day 14 embryos (by 28%). Several other studies confirmed lower pregnancy rates in endometriosis animals versus controls ( Barragan et al. , 1992 ; Stilley et al. , 2010 ; Sharpe-Timms et al. , 2020 ). The endometriotic lesions have been shown to regress in pregnant rats ( Vernon and Wilson, 1985 ; Rajkumar et al. , 1990 ) and during lactation demonstrating beneficial effects of anoestrous. This model also demonstrates anomalies of the hypothalamic–pituitary–ovarian axis, ovarian follicle development, and ovulatory dysfunction, including fewer ovarian follicles and corpora lutea with luteinized unruptured follicles ( Moon et al. , 1993 ; Stilley et al. , 2010 ). Furthermore, in vivo anomalies in postovulatory oocyte structure and preimplantation embryo development, including misaligned chromosomes, nuclear and cytoplasmic fragmentation, and delayed or arrested cleavage, as well as lower implantation rates and spontaneous abortions were found in this model ( Stilley et al. , 2009 ; Birt et al. , 2013b ; Sharpe-Timms et al. , 2020 ). Emerging evidence demonstrates transgenerational impacts, leading to reduced fertility in females and males developmentally exposed to endometriosis over two and three generations ( Birt et al. , 2013b , Sharpe-Timms et al. , 2020 ; Stilley et al. , 2009 ) A single fertility study in mice demonstrates that mice >13 weeks old with endometriosis had more resorbed foetuses, and the pups that were born were smaller than controls ( Elsherbini et al. , 2022 ). No significant differences in fertility were seen between control and endometriosis in mice <7 weeks old ( Elsherbini et al. , 2022 ). Given the current lack of studies and data exploring the impact of endometriosis on pregnancy outcomes in mice, our recommendations for fertility endpoints are to use the rat model of endometriosis and/or validate findings in mice. The time between disease induction and lesion harvesting in rodent lesions ranges from 24 h until 56 days with most studies examining lesions at 21 days ( Burns et al. , 2021 ). Early timepoints (<24 h) are used to study angiogenesis, neurogenesis, and immune responses of the lesions during the early attachment and remodelling phases ( Sanchez et al. , 2017 ; Burns et al. , 2021 ). Studies using unmodified endometrial tissue show lesion establishment at sites of attachment 24, 48, and 72 h after endometriosis initiation ( Kobayashi et al. , 2011 ; Burns et al. , 2018 ). Gene expression at these stages indicates increased immune, angiogenic, and remodelling activity, independent of oestrogen/ESR1 signalling ( Chen et al. , 2010 ; Burns et al. , 2018 ). By Day 7, lesion remodelling appears complete, with immune-related gene expression changes persisting at Days 7 and 14 ( Panir et al. , 2024 ). Beyond Day 7, studies focus on lesion growth, hormone responses, fibrosis, pain, infertility, and treatment efficacy ( Duan et al. , 2018 ; Fattori et al. , 2020 ; Fattori et al. , 2024 ). A common approach involves lesion growth for 3-4 weeks, followed by a 3–4-week treatment period to assess ( Soares et al. , 2012 ; Fattori et al. , 2020 ). For accurate treatment evaluation, lesions should persist for at least 6–8 weeks to confirm regression is due to therapy rather than natural resolution. Inbred mouse strains maintain stable lesion numbers post-induction, with limited new lesion formation ( Burns et al. , 2012 ; Jones et al. , 2018 ; Dorning et al. , 2021 ). Outbred strains develop fewer lesions, which regress over time. In humans, it has been postulated that endometriosis lesions may be able to seed new lesions, which could explain the recurrence of the disease post-hysterectomy ( Rizk et al. , 2014 ; Praetorius et al. , 2022 ). Similarly, in baboons, lesions maybe dynamic, with lesions disappearing and appearing again over time ( Harirchian et al. , 2012 ). Lesion size/volume has been shown to increase over time in the homologous mouse model ( Duan et al. , 2018 ), mainly due to oestrogen-driven epithelial proliferation ( Burns et al. , 2012 ; Liang et al. , 2018 ; Dorning et al. , 2021 ). Importantly, in the full thickness model, the lesions are cystic, and thus an increase in lesion volume can be contributed by an increase in cyst size with fluid secretion being a natural epithelial response to hormones and not to solid growth of the lesion tissue ( Kobayashi et al. , 2011 ; Dodds et al. , 2017 ). For pain behaviour assessment, metrics, and SOPs in rodent models of endometriosis, please refer to our EPHect companion paper ( Dodds et al. , 2025 ). Identifying suitable control groups to compare findings from experimental models of endometriosis is critical to ensure statistical analyses and data interpretation are valid. As a general practice for rigour and reproducibility, controls should hold all conditions constant except the independent variable. Several options have been used in past animal studies on endometriosis and, as with selecting an appropriate model, ‘best-fit’ controls are dependent on the hypothesis being tested. For example, studies may evaluate rodents ‘with endometriosis’ versus ‘no endometriosis’. If inducing endometriosis via the intraperitoneal injection method, accompanying controls often receive intraperitoneal injection of vehicle alone (e.g. Dodds et al. , 2019 ; Fattori et al. , 2020 ). As mentioned above, some researchers alternatively inject non-endometrial tissue (e.g. Somigliana et al. , 1999 ; Morris et al. , 2021 ). For surgical induction of endometriosis, control animals undergo sham surgical procedures (e.g. Burns et al. , 2018 ). Using the autologous model as an example, this includes adding sutures (alone or with non-endometrial tissue) to the intestinal mesentery and mechanical manipulation of the uterine horn (e.g. Torres-Reveron et al. , 2016 ; Castro et al. , 2021 ). In both scenarios, all animals undergo identical procedures, though the control group does not receive donor endometrial tissue. Other studies may examine an ‘intervention’ versus ‘no intervention’ in their model of endometriosis. In these cases, the independent variable is a treatment. Endometriosis is therefore induced via the same method for all animals, with an experimental group receiving the treatment (i.e. a drug) and controls a placebo (i.e. vehicle alone) (e.g. Jones et al. , 2018 ; Hirakawa et al. , 2022 ; Muraoka et al. , 2023 ). Tissue collection and processing are areas of standardization, but euthanasia is more dependent on local rules and regulations of local institutional, company, university, or government animal care and use protocols ( Hubrecht and Carter, 2019 ). Most investigators, however, euthanize rodents with carbon dioxide. A second method of euthanasia, such as cervical dislocation, may be required. For tissue collection, if animals are intact and cycling, we recommend euthanizing the rodents during the same stage of the oestrous cycle (i.e. oestrus) as genes/proteins in the lesions are often hormone responsive; therefore, ensuring a similar cycle stage will reduce unwanted variability due to hormonal differences (see Supplementary File S1 : EPHect-EM-Homologous SOP 8 and 14). With cycling animals, obtaining the uterine weight adds an additional layer of confirmation of the staging along with evaluating the cytology of a vaginal lavage. For studies using treatments, ovarian and other organ weights can also be important metrics to document changes due to treatment. Tissue processing can also be standardized (see Supplementary File S1 : EPHect-HM-Homologous SOP 7 and 13). Once the animal is euthanized, the first step is to obtain peritoneal cavity cells from a lavage. In the rat, peritoneal fluid may be sufficiently abundant to be collected without lavage. Second, once the animal is opened, blood should immediately be collected, either from the descending aorta or via cardiac puncture. Obtaining the blood quickly ensures that clotting does not occur but also allows for a cleaner environment for lesion visualization, imaging, and descriptions (see Supplementary File S1 : EPHect-EM-Homologous SOP 10). Rapid recovery of tissues is required, and tissues need to be stored on wet ice until they have been weighed, measured, and documented. When wanting to study gene expression, at least one lesion from each animal should be frozen for nucleic acid assessment and a second fixed for histology to confirm lesion structures. Lesions can be kept for DNA, RNA, and protein isolation by snap freezing in liquid nitrogen or on dry ice. Technologies for nucleic acid analysis (both RNA and DNA) from fixed tissues are becoming more commonplace but still require careful selection of methods, training, and controls to avoid artefactual data ( Yong et al. , 2021 ). Fixing the lesions by formalin or paraformaldehyde allows for superior histological analysis to visualize glands, stroma, immune infiltrates, and fibrosis (compared to observations in frozen tissues). For fixed lesions, we recommend the collection of tissue adjacent to the lesion, which will allow visualization of invasion of surrounding tissues and whether adhesions are present. When freezing the lesions, we recommend removing extra tissues or planning for cell-type enrichment with microdissection to prevent contamination of non-lesion tissue in DNA, RNA, or protein results. For the other organs (e.g. kidney, spleen, etc.), if there are two, one can be fixed and one can be frozen. Alternatively, a small piece of organ or lesion, if large enough, can be cut off for fixing, with the rest being snap frozen.

Our knowledge of endometriosis and endometriosis-related animal models has advanced significantly in the last decade. However, differences in experimental design, execution, and analysis of studies using experimental models of endometriosis have contributed to discrepancies in the literature, resulting in an overall inability to replicate and validate research results. Having identified the unmet need to develop standards of practice in experimental models of endometriosis, the EPHect team convened a working group led by experts in rodent models for endometriosis. The goal of the working group was to develop standardized resources and recommendations to be used in endometriosis research when deploying experimental models to improve continuity, reduce variability, and ensure studies are properly controlled. This will facilitate comparisons across laboratory groups, as well as enable multi-centre collaborations. The resulting recommendations are deliberately kept open to ensure hypothesis-driven discovery-based science would not be hampered in the absence of a defined aetiology. While standardization is advocated, we acknowledge that most models we evaluated are ‘generic’. This standardization, therefore, should not be construed to stifle innovation in establishing new models of endometriosis. We advocate that with each novel model developed, the recommended standardized documentation and reporting details are embraced to facilitate result dissemination, inference, and comparison. There are and will be rodent models of endometriosis that may better represent specific phenotypes and that can explore specific outcomes. For example, to investigate specifically, the effect of ovarian endometrioma on fertility, a targeted mouse model has been reported ( Hayashi et al. , 2020 ). Derived from the advantages and disadvantages of each model, a decision tree was made to formalize our recommendations ( Fig. 3 ). Our recommendations are also supported by the work of Nunez-Badinez et al. (2021) , who found that currently, the most common homologous mouse model of endometriosis is the syngeneic intraperitoneal injection model. They also found that a high proportion of papers published used the autologous suture model of endometriosis, but this mouse model is no longer commonly used by laboratories ( Nunez-Badinez et al. , 2021 ). Thus, like their findings, we too endorse the use of endometrium injected into the peritoneal cavity (wherever possible in ovary-intact animals) to develop endometriosis-like lesions in the mouse. The rat model, on the other hand, does not form robust lesions with intraperitoneal injection of uterine endometrium ( Persoons et al. , 2020 ); therefore, the suture of size-defined full-thickness uterine tissue is recommended in the rat model. Decision tree of key variables in homologous murine models of endometriosis. The five main variables identified are shown in blue. Recommendations for mouse and rat are shown in green. IP, intraperitoneal; CD-1/ICR, Hsd: ICR (CD-19 ® ); ddY, DDY/JcISidSeyFrkJ; FVB/N, Friend virus B NIH. Based on the myriad of commonalities among protocols received, we developed multiple SOPs to complement these recommendations for mice regarding pre- and post-operative care ( Supplementary File S1 : EPHect-EM-Homologous SOPs 1 and 10, respectively), injection of hormones ( Supplementary File S1 : EPHect-EM-Homologous SOPs 2 and 3), decidualization ( Supplementary File S1 : EPHect-EM-Homologous SOP 4), injection procedures ( Supplementary File S1 : EPHect-EM-Homologous SOPs 5 and 6), endpoint assessment and lesion collection ( Supplementary File S1 : EPHect-EM-Homologous-SOP 7), oestrous cycle staging ( Supplementary File S1 : EPHect-EM-Homologous SOP 8), and ovariectomy ( Supplementary File S1 : EPHect-EM-Homologous SOP 9). In a similar manner, SOPs were created for the rat model of endometriosis regarding pre- and post-operative care ( Supplementary File S1 : EPHect-EM-Homologous SOPs 11 and 15, respectively), induction of endometriosis ( Supplementary File S1 : EPHect-EM-Homologous SOP 12), endpoint assessment and lesion recovery ( Supplementary File S1 : EPHect-EM-Homologous SOP 13), and oestrous cycle staging ( Supplementary File S1 : EPHect-EM-Homologous SOP 14). In addition to the SOPs, to reduce variability, we also recommend detailed laboratory notebook keeping to record lesion location, volume, and colour, and other discernible characteristics ( Fig. 2 ). In mice, we recommend at minimum that inbred mouse strains are used, uterine tissue is engrafted within the peritoneal cavity at a ratio of one donor mouse to two recipient mice, and exposure to oestrogens. Our standard recommendations are that C57BL/6 mice are used when possible, and that uterine tissue is engrafted via intra-peritoneal injection at a ratio of 1:1 in ovary-intact recipients (see Minimal and Standard Recommendations in Table 3 ). We envision that this will ensure data rigour, reproducibility, and repeatability from group to group. Recommendations are more simplistic for rat models, as less variation exists, but equally important to maintain replicability moving forward. Our standard recommendations are to auto-transplant 4–6 uterine tissue pieces onto the arterial cascades of the small intestine of ovary-intact Sprague-Dawley rats. We also recommend collecting a full repertoire of biospecimens and endpoint assessments in mice and rats to enable future collaborations or address new hypotheses whilst reducing the number of animals used in research. In studies that evaluate potential therapeutics, or aim to discover new therapeutic targets or biomarkers, we highly recommend the use of multiple pre-clinical models. This will test the accuracy/rigour of the discovery science in models that may represent different aetiologies. Regardless of the model decision made by individual groups, it will be imperative for both the research and peer reviewer community to encourage use of these standards and, at the very least, require complete and transparent reporting on the core variables described above. The resources and recommendations we have developed will aid in cross-institute collaboration for this exact purpose. EPHect-EM-homologous standard operating procedures: minimum and standard documentation for homologous rodent models. We anticipate this EPHect initiative will inspire a new chapter in the standardization of endometriosis rodent model studies, including standardization of collection methods. Our recommendations on rodent models of endometriosis will inspire and foster new collaborations among existing endometriosis laboratory groups and encourage other investigators to join the field. This will allow us, as a community of researchers, to aid in the journey to understand the pathogenesis/aetiology of endometriosis that will, ultimately, be translated into improved treatments for the millions of those affected by this disease worldwide.