A retrospective study of Pipelle endometrial biopsy for ovarian, fallopian tube, and peritoneal cancers

Abstract Objectives Although the Pipelle endometrial biopsy is widely performed as a practical and minimally invasive test for endometrial disease(s), its effectiveness in ovarian cancer has not been explored. The aim of the present study was to evaluate the results of Pipelle endometrial biopsy for ovarian, fallopian tube, and peritoneal cancers. Methods A pre-treatment Pipelle-endometrial biopsy was performed in 90 patients with ovarian, fallopian tube, or peritoneal cancers between January 2014 and November 2021. We retrospectively analysed the association between the results of Pipelle endometrial biopsy and clinicopathological data. Moreover, we evaluated their impact on the following treatment in advanced cases initially treated with chemotherapy. Results The sensitivity and false-negative rates for Pipelle endometrial biopsy were 25/90 (27.8%) and 65/90 (72.2%) in all patients, respectively, and 23/56 (41.0%) and 33/56 (58.9%) in cases with advanced disease (stages III and IV), respectively. Pipelle-positive endometrial biopsy-positive (Pipelle-positive) was not observed in 29 patients with clinical stage I disease, and Pipelle-positive patients exhibited significantly more high-grade serous carcinomas, and positive peritoneal, endometrial, and cervical cytologies than Pipelle-endometrial biopsy-negative cases. Surgical pathology was confirmed in 23 Pipelle-positive patients, and 17/23 (74.0%) had the same diagnosis as that for Pipelle endometrial biopsy. Conversely, 6/23 (26.0%) patients exhibited a minor diagnostic discrepancy between Pipelle endometrial biopsy and surgical pathology. Nineteen of the 38 (50.0%) patients initially treated with chemotherapy were identified as Pipelle-positive, contributing to a prompt histological diagnosis and pre-treatment tumour sampling. Companion diagnostic tests were performed using Pipelle endometrial biopsy samples from 4 inoperable patients. Conclusion Although the positive rate of Pipelle endometrial biopsy in ovarian, fallopian tube, and peritoneal cancers is low, Pipelle endometrial biopsy may enable prompt histological diagnosis and initiation of chemotherapy while collecting tumour tissue for genetic testing in some cases with advanced disease. PLAIN LANGUAGE SUMMARY The effectiveness of pre-treatment Pipelle endometrial biopsy for ovarian, fallopian tube, and peritoneal cancers remains unclear. This study demonstrated that Pipelle endometrial biopsy may enable prompt histological diagnosis and initiation of chemotherapy while collecting tumour tissue for genetic testing in some cases with advanced disease. This was a single-centre, retrospective study; as such, the effectiveness of Pipelle endometrial biopsy should be evaluated in larger prospective studies, including comparisons with other tumour sampling methods.


Introduction
More than one-half of ovarian, fallopian tube, and peritoneal cancers are diagnosed at an advanced stage (Park et al. 2017).Previous prospective randomised trials have demonstrated that neoadjuvant chemotherapy (NAC), followed by interval debulking surgery, is an effective treatment for poor surgical candidates or patients with a low likelihood of optimal cytoreduction (Vergote et al. 2010, Kehoe et al. 2015, Onda et al. 2016).In NAC cases, pre-treatment tumour sampling using minimally invasive techniques, followed by prompt initiation of chemotherapy, is preferred (Wright et al. 2016, Armstrong et al. 2021).However, pre-treatment tumour sampling is not feasible in some cases, and cytopathology from ascites or pleural effusion combined with tumour markers can be used before treatment in these cases (Vergote et al. 2010, Kehoe et al. 2015, Onda et al. 2016).
There have been several reports of ovarian cancer cells detected by endometrial or cervical cytology (Hirasawa et al. 1997, Takeshima et al. 1997, Jobo et al. 1999, Sasagawa et al. 2003, Suzuki et al. 2010).These findings are believed to represent the appearance of ovarian cancer cells in the uterine cavity through the fallopian tubes.Takashina et al. reported that the positivity rates for endometrial and cervical cytology in ovarian cancer were 41.9% and 19.3%, respectively (Takashina et al. 1988).Positive endometrial cytology is common in high-grade serous carcinoma (HGSC), clear cell carcinoma (CCC), ascitic-positive cases, and in cases with advanced clinical stage (Hirasawa et al. 1997, Jobo et al. 1999).Although several reports have demonstrated the utility of endometrial cytology in the early detection of ovarian, tubal, and peritoneal cancers (Maeda et al. 2010, Otsuka et al. 2013), its clinical utility in ovarian cancer is limited, especially in advanced cases.
Recently, Pipelle endometrial biopsies have been widely performed as an effective, minimally invasive, and low-cost test for endometrial diseases (Stovall et al. 1991, Terzic et al. 2020, Kaiyrlykyzy et al. 2021).We actively perform Pipelle endometrial biopsy as a preoperative test to detect endometrial cancer in patients with suspected ovarian, fallopian tube, or peritoneal cancers.We found that cancerous tissues were collected in some patients without synchronous endometrial cancer.However, the effectiveness of Pipelle endometrial biopsy for ovarian, fallopian tube, and peritoneal cancers remains unclear.
In the present study, 90 patients with ovarian, fallopian tube, and peritoneal cancers underwent pre-treatment Pipelle endometrial biopsy.We retrospectively analysed the association between the results of Pipelle endometrial biopsy and clinicopathological data.Moreover, we evaluated their impact on the following treatment in advanced cases initially treated with chemotherapy.

Clinical data analysis
The study has been reported in line with the strengthening of the reporting of observational studies in epidemiology STROBE criteria (von Elm et al. 2007).A total of 325 patients with ovarian, fallopian tube, and peritoneal cancers treated at Niigata Cancer Centre Hospital (Niigata, Japan), between January 2014 and November 2021 were included.Of these, 37 cases (20 synchronous endometrial abnormalities [19 endometrial cancer and 1 atypical endometrial hyperplasia] and 17 post-hysterectomy) were excluded.None of the patients had synchronous cervical cancers.Of the 288 curated cases, 90 patients who underwent pre-treatment Pipelle endometrial biopsy in an outpatient setting were retrospectively analysed (Figure 1).For Pipelle endometrial biopsy, samples were obtained from the uterine cavity using a pipette (Medgyn IV, Harada Corporation, Tokyo, Japan) or endosuction (open end type; Hakko, Nagano, Japan).To evaluate sample selection bias, the clinicopathological characteristics of patients who underwent biopsy were compared with those who did not.As a result, more advanced cases of HGSC and fewer patients with CCC underwent Pipelle endometrial biopsy (Supplementary Table 1).During screening, all patients underwent clinical assessment, computed tomography (CT), magnetic resonance imaging (MRI), or both, and serum tumour marker (carbohydrate antigen 125 [CA125] and carcinoembryonic antigen [CEA]) measurement before treatment.
Clinical findings (age, clinical stage, serum tumour marker level, CT/MRI, and treatment course), pathological findings (Pipelle endometrial biopsy, surgical pathology, and cytology), and results of companion diagnostic tests (MyChoice CDx, Myriad Genetic Laboratories, Inc., Salt Lake City, USA) and the microsatellite instability [MSI] test [SRL, Tokyo, Japan]) were retrospectively analysed.A pathological review was performed by gynecological pathologists; surgical pathology was also reviewed in Pipelle-positive cases.The tumours were classified according to the 2020 World Health Organisation classification system (version 5).In this study, a positive Pipelle endometrial biopsy result was defined as the detection of adenocarcinoma, pathologically considered to be ovarian, fallopian tube, or peritoneal cancer(s).All patients were evaluated using immunostaining for Wilms Tumour 1 (WT1) and paired box 8 (PAX8), which is considered to be effective in diagnosing ovarian cancer (Liliac et al. 2013).The results of peritoneal, endometrial, and cervical cytologies performed at the authors' hospital or other referral hospitals were also investigated.A positive result on peritoneal and endometrial cytology was defined as suspicion of malignancy.A positive cervical cytology result was defined as adenocarcinoma.

Statistical analysis
All analyses were performed using R statistics (R Core Team [2018], language, and environment for statistical computing [R Foundation for Statistical Computing, Vienna, Austria; http://www.R-project.org/]).Categorical variables were compared between two groups using Fisher's exact test (Table 1, Figure 2b, and Supplementary Table 1), and continuous variables among the groups were compared using the Kruskal-Wallis test with the Steel-Dwass post-hoc test (Figure 2c).Differences with a two-tailed p < 0.05 were considered to be statistically significant.

Details of 25 Pipelle endometrial biopsy-positive patients
Histology of the 25 Pipelle-positive cases revealed 16 HGSCs, 4 mixed histological tumours (HGSC and CCC), 3 CCCs, and 2 serous carcinomas (grade undetermined).The median tumour size and content of the specimens on the slide were 25 (range, 1-225) mm 2 and 30% (range, 1%-60%), respectively.All 25 cases were positive for PAX8, and WT1 was positive in 21/25 (84%).Following Pipelle endometrial biopsy, 19/25 (76.0%) patients underwent chemotherapy as the initial treatment.In 23 patients, surgical pathology results were confirmed, and 17/23 (74.0%) had a diagnosis completely similar to that of Pipelle endometrial biopsy.Conversely, 6/23 (26.0%) patients exhibited a diagnostic discrepancy between the Pipelle endometrial biopsy and surgical pathology results.Two serous carcinoma cases (undetermined grade) in the biopsy samples were diagnosed as HGSC based on surgical pathology (cases 76 and 79).Of the 2 cases of mixed tumours (i.e., HGSC þ CCC) in Pipelle endometrial biopsy, one was diagnosed as HGSC and the other as CCC based on surgical specimens.In the 2 cases surgically diagnosed with carcinosarcoma (case 59, HGSC, endometrial stromal sarcoma,  and chondrosarcoma; case 83, CCC, chondrosarcoma, and undifferentiated sarcoma), the carcinoma component-but not the sarcoma component-was accurately detected in the biopsy sample.Hysterectomy was performed in 20/25 (80%) Pipelle-positive cases and no case of endometrial cancer was suspected.Although there were 3 cases of suspected endometrial invasion on MRI, the surgical specimens were negative for endometrial cancer in two cases (1 and 66).Surgical pathology was unavailable in 1 patient (case 44); however, WT-1 was positive on immunohistochemical analysis, suggesting that the tumour originated from the ovarian and fallopian tubes (Table 2 and Supplementary Table 3).

The impact of Pipelle endometrial biopsy for patients initially treated with chemotherapy
Data from 38 patients, who were initially treated with chemotherapy, were used to evaluate the effect of Pipelle   2a and Supplementary Table 4).Significantly fewer Pipelle-positive patients underwent diagnostic surgery than those who were Pipelle-negative (Figure 2b).The number of days from the first visit to diagnosis in the Pipelle endometrial biopsy group was comparable to that in the cell block diagnosis group, and significantly shorter than that in the diagnostic laparoscopy group (median 7 days versus [vs.] 7 days vs. 19 days; p < 0.001) (Figure 2c).Using samples collected from Pipelle endometrial biopsies, 4 companion diagnostic tests (3 myChoice CDx tests and 1 MSI test) were performed on four inoperable HGSCs (cases 44, 66, 76, and 86).All three patients who underwent myChoice CDx were homologous recombination deficiency (HRD)-positive, and maintenance therapy with poly (ADP-ribose) polymerase (PARP) inhibitors was performed (Figure 3).
HGSCs are believed to originate from the fallopian tubes (Kurman andShih Ie, 2010, Labidi-Galy et al. 2017).In addition to HGSC, 3/12 (25%) CCCs were Pipelle-positive, all of which were clinical stage II or higher, with positive peritoneal cytology.This suggests that, in some advanced cases of both HGSC and non-HGSCs, there is isolated tumour tissue in the uterine cavity and fallopian tubes, and the negative pressure applied in Pipelle endometrial biopsy (Du et al. 2016) may collect the isolated tumour tissue.Pipelle endometrial biopsy can be performed easily in an outpatient setting at the time of the first visit and merits and attempt considering the advantages of early histological diagnosis and specimen collection, especially in advanced cases.Importantly, 19/38 (50%) advanced cases initially treated with chemotherapy were Pipelle-positive.In NAC cases, histological diagnosis using pre-treatment biopsy specimens is preferred because cytology may be insufficient to distinguish between borderline and invasive cancers, and tumour cells may be too necrotic to be identified after exposure to chemotherapy (Wright et al. 2016, Onda et al. 2021).However, preoperative histological diagnosis was difficult to establish in many cases.

PIPELLE BIOSPY FOR OVARIAN CANCER
the Primary chemotherapy versus primary surgery for newly diagnosed advanced ovarian cancer (CHORUS): an open-label, randomised, controlled, non-inferiority trial were diagnosed only by cytology.Diagnostic laparoscopy is increasingly performed because it is minimally invasive and can be effective in staging to predict the results of primary cytoreductive surgery and tumour biopsy (Fagotti et al. 2006, Fagotti et al. 2016, Rutten et al. 2017).However, diagnostic laparoscopy is performed under general anaesthesia, which is accompanied by the risk for postoperative complications and increased costs (Kim et al. 2021).Moreover, treatment delays have been associated with poor outcomes in patients with ovarian cancer (Tewari et al. 2016, Timmermans et al. 2018).Therefore, indications for diagnostic surgery should be carefully assessed.Our study suggests that in advanced cases that are difficult to diagnose, Pipelle endometrial biopsy may enable early histological diagnosis in a minimally invasive manner.
We also demonstrated that Pipelle endometrial biopsy can collect a sufficient volume of specimens for companion diagnostic tests in some advanced cases.Although several recent studies have reported the possibility of detecting ovarian cancer using the liquid obtained from the Papanicolaou test and uterine lavage to analyse genetic mutations, the low sensitivity of these tests is considered to be a vital issue (Kinde et al. 2013, Maritschnegg et al. 2015, Wang et al. 2018, van Bommel et al. 2022).The identification of HRD is critical for selecting patients eligible for treatment with PARP inhibitors, and MyChoice CDx is now considered to be the standard first-line therapy for ovarian, fallopian tube, and peritoneal cancers (Miller et al. 2020, Armstrong et al. 2022).Pre-treatment tumour samples are preferred for MyChoice CDx because post-NAC samples have problems including a high frequency of test failures owing to low tumour DNA percentages and possible discrepancies with pre-treatment tumour samples (Takaya et al. 2020, Zalaznick et al. 2022).Diagnostic laparoscopy and exploratory laparotomy are the most reliable methods for collecting tumour samples; however, surgery is difficult to perform in some cases.In this study, 19/38 (50.0%) patients initially treated with chemotherapy were Pipelle-positive.Moreover, significantly fewer Pipelle-positive patients underwent diagnostic surgery than Pipelle-negative patients.This may be due to the avoidance of surgery to obtain tumour tissue for histological diagnosis and companion diagnostic tests in Pipelle-positive cases, in which upfront surgery is difficult to perform based on imaging findings or a poor general condition.Cell block diagnosis of ascites cells, another minimally invasive method that can potentially obtain specimens for companion diagnostic tests in a short time, was performed in of 16/38 (42.1%) advanced cases initially treated with chemotherapy.However, companion diagnostic tests using cytology samples are often challenging due to the low quality and quantity of samples, and the difficulty of preparation (Roy-Chowdhuri et al. 2015, Balla et al. 2018, Pisapia et al. 2021); thus, it is not a reliable method at present.Moreover, the collection of ascites is difficult in some advanced cases.Although Pipelle endometrial biopsy is not a substitute for diagnostic surgery because it does not offer the possibility of intraperitoneal observation, our findings suggest that it is a minimally invasive tumour sampling method for genetic testing in some primarily inoperable patients.
In contrast, the false-negative rate of Pipelle endometrial biopsy was high and 65/90 (72.2%) patients were Pipellenegative.In particular, the utility of Pipelle endometrial biopsy in stage I ovarian, fallopian tube, and peritoneal cancers has not been confirmed.There are two possible reasons for this.First, there may be no isolated tumour tissues in the fallopian tubes or abdominal cavity in stage I cases, except for fallopian tube cancer.The other is the possibility of sampling error.A problem with Pipelle endometrial biopsy is insufficient specimen collection (Terzic et al. 2020, Kaiyrlykyzy et al. 2021), especially in postmenopausal women with a thinning endometrium (Elsandabesee and Greenwood, 2005).Consistent with the possibility of sampling error, 16/65 (24.6%)Pipelle-negative cases had an insufficient sample volume in this study.Therefore, less invasive endometrial cytology may be more useful than Pipelle endometrial biopsy as a screening test for endometrial diseases in early cases.However, because Pipelle-positive samples are infrequent, even in advanced cases, another tumour sampling method should be considered simultaneously in advanced cases.
The present investigation was a single-centre retrospective study; as such, the utility of Pipelle endometrial biopsy should be evaluated in larger prospective studies, including comparisons with other tumour sampling methods.Furthermore, there were other important limitations to this study.First, it is difficult to completely rule out the presence of endometrial cancer when the Pipelle endometrial test result is positive.Hysterectomy was performed in 20/25 (80.0%)Pipelle-positive cases and no case of endometrial cancer was suspected.However, 15/20 (75.0%) patients underwent hysterectomy after NAC, and the possibility that the endometrial cancer had disappeared could not be excluded.Next, in 6/23 (26.0%) cases in which surgery was performed, a diagnostic discrepancy between Pipelle endometrial biopsy and surgical pathology was observed.In 4 cases, the tumour component could not be accurately assessed (cases 20, 21, 59, and 83).Four patients had small specimens, with tumour size, tumour content-or both-far below the median (Table 2).Accurate histological tumour grading was also difficult for the 2 HGSCs in the Pipelle endometrial biopsy samples.Finally, we performed a companion diagnostic test using Pipelle endometrial biopsy samples from only 4 cases in this study.For the MyChoice CDx and MSI tests, 30% and 50% tumour content are ideally required, respectively (Patel et al. 2018, Shimozaki et al. 2021).Among the specimens collected in this study, 14/25 (56.0%) and 7/25 (28.0%) met the criteria; however, it may be difficult to perform these tests on other specimens.
In conclusion, the positive rate of Pipelle endometrial biopsy in ovarian, fallopian tube, and peritoneal cancers is low and another tumour sampling method should be considered simultaneously.Pipelle endometrial biopsy may enable prompt histological diagnosis and initiation of chemotherapy while collecting tumour tissues for genetic testing in some primarily inoperable patients.

Figure 2 .
Figure 2. Pre-treatment tumour sampling methods in 38 patients with advanced ovarian cancer, fallopian tube cancer, and peritoneal cancer primarily treated with chemotherapy.(a) Other tumour sampling methods in the Pipelle endometrial biopsy-positive and negative groups are shown.The squares represent the respective cases, and the colours indicate the sampling method.(b) The bar chart illustrates the difference between Pipelle-positive and Pipelle-negative patients with and without diagnostic surgery.(c) Boxplots depicting the days from the first visit to diagnosis for each tumour sampling method.

Figure 3 .
Figure 3. Four primarily inoperable patients with high-grade serous carcinoma (HGSC) who underwent companion diagnostic tests using samples collected via Pipelle endometrial biopsy.Representative images of haematoxylin and eosin staining of the results of companion diagnostic tests in 4 cases are shown (original magnification, �100 and �400).

Table 1 .
Clinicopathological characteristics of 90 patients who underwent Pipelle endometrial biopsy before treatment.
(Yedema et al. 1992)n diagnosis and treatment.All 38 patients initially treated with chemotherapy were suspected to have stage III or IV ovarian, fallopian tube, or peritoneal cancer on CT/MRI, and the CA125/CEA ratio was > 25 in all except for 1, which is useful for distinguishing ovarian cancer from colorectal adenocarcinoma(Yedema et al. 1992).Details of the clinicopathological findings in the 38 patients initially treated with chemotherapy are summarised in Supplementary Table3.Nineteen of the 38 (50.0%) patients tested positive for Pittsburgh compound B (PiB).For pre-treatment tumour sampling, 32 (84.2%) patients underwent tumour sampling in conjunction with other sampling methods.In particular, 14/38 (36.8%) patients underwent diagnostic surgery (12 diagnostic laparoscopies and 2 exploratory laparotomies) and 16/38 (42.1%) patients underwent cell block diagnosis of ascites or pleural fluid.In contrast, 6/38 (15.8%) patients had a pre-treatment histological diagnosis based only on Pipelle endometrial biopsy (Figure