Intraoperative pulmonary air-leak sites in catamenial pneumothorax: a retrospective 15-case surgical series

Catamenial pneumothorax is a distinct form of pneumothorax that occurs in women of reproductive age. It is a major clinical manifestation of thoracic endometriosis. In a nationwide Japanese database survey of 157,087 patients hospitalized for pneumothorax, female patients comprised 17.6% of the cases, and catamenial pneumothorax accounted for 3.1% of all female cases and 7.5% of cases among women aged 13–53 years ( 1 ). In contrast, surgical series focusing on women of reproductive age have reported substantially higher proportions, ranging from approximately 20% to 35% ( 2 - 4 ), suggesting that catamenial pneumothorax may be underrecognized in routine clinical practice. The pathogenesis of catamenial pneumothorax remains unclear, and several mechanisms have been proposed, including transdiaphragmatic air passage through diaphragmatic fenestrations, migration or dissemination of endometrial tissue, coelomic metaplasia, and rupture of visceral pleural or subpleural endometriotic lesions resulting in air leakage ( 5 ). Diaphragmatic involvement appears to be the dominant thoracic lesion in surgically treated patients; previous surgical series reported diaphragmatic lesions in 96% of patients with histologically confirmed thoracic endometriosis and in all patients with surgico-pathologically confirmed thoracic endometriosis-related pneumothorax ( 3 , 4 ). Nevertheless, endometriosis-related pneumothorax may occur outside the menstrual period, and visceral pleural endometriotic implants can be present without evident diaphragmatic defects ( 2 , 3 ), suggesting that visceral pleural lesions may contribute directly to air leakage in selected cases. A previous study described the characteristic distribution of visceral pleural endometriotic lesions, particularly in segment S4 and the three-lobe confluence ( 6 ). In the large series by Tsuboshima et al. , pulmonary air-leak sites were also documented in a limited subset of patients with prolonged preoperative air leakage; however, among 248 surgically treated patients, detailed air-leak site findings were available for only 13 patients ( 7 ). Therefore, although the anatomical distribution of pulmonary air-leak sites has been reported, several important aspects remain unclear, including the frequency of intraoperative pulmonary air-leak detection in consecutive pathologically confirmed cases, the pathological correlation between air-leak sites and endometriotic lesions, and the relationship between leak-site identification and postoperative recurrence. In recent years, our surgical approach has evolved to include more systematic intraoperative air-leak testing and thorough inspection of the entire lung surface, alongside routine assessment of diaphragmatic lesions. During this process, we repeatedly encountered patients with intraoperative pulmonary air leakage, particularly in the middle lobe. Therefore, we retrospectively reviewed surgically treated patients with pathologically confirmed catamenial pneumothorax to evaluate the frequency and anatomical location of intraoperative pulmonary air leaks, characterize the pathological features of resected air-leak lesions, describe the surgical procedures performed at Yokohama Municipal Citizen’s Hospital, and evaluate postoperative recurrence according to whether a pulmonary air-leak site was identified intraoperatively. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1411/rc ) ( 8 ).

We retrospectively reviewed the data of all female patients who underwent surgery for pneumothorax at Yokohama Municipal Citizen’s Hospital between October 2008 and December 2025. Patients with histopathologically confirmed thoracic endometriosis in resected diaphragmatic and/or pulmonary specimens were included in this study. The diagnosis was supported by immunohistochemical findings, including estrogen receptor and/or CD10 positivity. The following clinicopathological features were retrospectively obtained from medical records: age at surgery, side of pneumothorax, smoking history, history of pelvic endometriosis, preoperative hormonal therapy, association between pneumothorax and menstruation, intraoperative findings of diaphragmatic fenestrations, presence and location of intraoperative air leaks, surgical procedures performed, pathological findings, postoperative hormonal therapy, postoperative recurrence, treatment for recurrence, and follow-up period. Surgical management was not standardized during the study period. Decisions regarding pulmonary wedge resection in the absence of an identifiable air leak and the need for diaphragmatic treatment or reinforcement were made according to intraoperative findings and the surgeon’s judgment. Pleurodesis was not performed in this cohort. Pathological evaluation of the resected specimens was performed using hematoxylin and eosin staining, along with immunohistochemical examination for estrogen receptor and/or CD10. Thoracic endometriosis was diagnosed when endometrial stromal and/or glandular components were identified in resected specimens. The site responsible for the pneumothorax was retrospectively assessed based on operative records and surgical videos, when available. Particular attention was paid to the site of directly observed intraoperative air leakage and the presence of diaphragmatic fenestrations. Postoperative recurrence was defined as an ipsilateral pneumothorax diagnosed radiologically during the follow-up period after surgery. The follow-up period was calculated from the date of surgery to the last follow-up visit. This study was approved by the Institutional Review Board of the Yokohama Municipal Citizen’s Hospital (approval No. Z260410). The requirement for written informed consent was waived because all patient data were analyzed anonymously. This study was conducted in accordance with the principles of the Declaration of Helsinki and its subsequent amendments. Because of the small sample size, statistical analyses were primarily descriptive, and no multivariable analysis was performed. Recurrence-free survival was estimated using the Kaplan-Meier method. All analyses were performed using IBM SPSS Statistics (version 29; IBM Corp., Armonk, NY, USA).

During the study period, 76 female patients underwent surgery for pneumothorax. Of these, 15 patients with pathologically confirmed thoracic endometriosis were included in the study cohort. The median age at surgery was 42 years (range, 27–54 years). Five patients had a history of pelvic endometriosis, three had received preoperative dienogest, and nine developed pneumothorax during menstruation. All 15 patients had a right-sided pneumothorax. Surgery was performed during the first episode of pneumothorax in two patients, whereas the remaining 13 underwent surgery for recurrent pneumothorax. The indications for surgery included persistent air leakage despite preoperative chest tube drainage in nine patients, radiological progression of pneumothorax in three patients, and preventive surgery in three patients ( Table 1 ). Among the 12 patients who underwent surgery with either active or suspected ongoing air leakage, an intraoperative pulmonary air-leak site was identified in nine patients. Macroscopically identifiable diaphragmatic lesions were present in all patients during surgery and were categorized as diaphragmatic defects or areas of thinning ( Figure 1A ). The leak site was most commonly located in segment S4 (six patients; Figure 1B,1C ), followed by segments S1, S6, and S8 (one patient each). When pulmonary air leakage was identified, wedge resection of the corresponding lesion was performed. In one patient with an S8 air-leak site, an additional S4 bullous lesion was also resected. In addition, two patients without detectable pulmonary air leakage underwent wedge resection of bullous lesions. Management of diaphragmatic lesions varied according to the intraoperative findings and the surgeon’s judgment and included ligation, resection, coverage, or no direct diaphragmatic procedure. Intraoperative findings in catamenial pneumothorax. (A) Intraoperative observations of diaphragmatic lesions, including defects and thinning. (B) Thoracoscopic view of the confluence of the three lobes. The arrow indicates a bulla on the middle lobe near the interlobar surface. (C) Air leakage from the bulla confirmed by intraoperative leak testing. Pathological examination revealed endometrial lesions in all 12 resected pulmonary specimens ( Figure 2 ) and in seven of eight resected diaphragmatic specimens ( Figure 3 ). Notably, endometrial lesions were confirmed in all pulmonary specimens obtained from an intraoperatively identified air-leak site. Although diaphragmatic lesions were identified macroscopically in all patients, diaphragmatic tissue was not obtained from seven patients. One patient with macroscopic diaphragmatic thinning had no histological evidence of endometriosis in the resected diaphragmatic specimen. The detailed intraoperative findings, surgical procedures, and pathological findings are summarized in Table 2 . Pathological findings of the pulmonary lesion responsible for air leakage. (A) Macroscopic view of a resected pulmonary specimen. The arrow indicates the bulla corresponding to the intraoperative air-leak site. (B) Microscopic view of endometrial tissue in the pulmonary lesion (hematoxylin and eosin stain at ×40). (C) Microscopic view of endometrial tissue in the pulmonary lesion (hematoxylin and eosin stain at ×400). (D) Immunohistochemical staining demonstrating estrogen receptor positivity in the corresponding area (×400). Pathological findings of the diaphragmatic lesion. (A) Macroscopic view of a resected diaphragmatic specimen. The arrow indicates a diaphragmatic defect. (B) Microscopic view of endometrial tissue in the diaphragmatic lesion (hematoxylin and eosin stain at ×400). Immunohistochemical staining showing positivity for (C) estrogen receptor and (D) CD10 in the corresponding lesion (×400). † , endometrial lesions were histopathologically confirmed. PGA, polyglycolic acid. During follow-up, five of the 15 patients (33.3%) experienced postoperative recurrence. For the entire cohort, the median follow-up period was 18.0 months (range, 2.5–136.5 months). Kaplan-Meier analysis demonstrated 1- and 2-year recurrence-free survival rates of 85.6% and 63.4%, respectively. Recurrence was observed in 1 of 9 patients (11.1%) with an intraoperatively identified leakage site, compared with 4 of 6 patients (66.7%) without an identified leakage site. Postoperative hormonal therapy, recurrence, follow-up duration, and treatment for recurrence are summarized in Table 3 . All patients with recurrence had received postoperative hormonal therapy; however, in one patient, recurrence developed after the completion of hormonal treatment. In all recurrent cases, the pneumothorax was mild and managed conservatively by continuation of hormonal therapy or a change in hormonal agent, without chest drainage or reoperation.

In this study, pulmonary air leaks were identified intraoperatively in 60% of surgically treated patients with catamenial pneumothorax and were most frequently located in the middle lobe, particularly in segment S4 near the confluence of the three lobes. Although diaphragmatic abnormalities are widely recognized as characteristic features of catamenial pneumothorax, our findings suggest that pulmonary lesions may directly contribute to the development of pneumothorax. This distinction is clinically important because it indicates that intraoperative evaluation should focus on both diaphragmatic lesions and thorough inspection of the middle lobe and its adjacent interlobar surfaces. These findings were largely consistent with the distribution of visceral pleural endometriotic lesions reported by Ochi et al. ( 6 ). They reported that segment S4 accounted for 63.5% of the 104 visceral pleural lesions, followed by segment S6 (12.5%), segment S2 (6.7%), segment S1 (5.8%), the basal segment (5.8%), segment S5 (3.8%), and segment S3 (1.9%). Importantly, 82.7% of all lesions were concentrated in the junctions of the upper, middle, and lower lobes. Based on this characteristic distribution, they proposed a “shelter effect” hypothesis, whereby thoracic endometriotic lesions preferentially implant in relatively stable, protected regions of the pleural cavity. Tsuboshima et al. extended these observations by reviewing 248 surgically treated patients with thoracic endometriosis-related pneumothorax and identifying intraoperative air-leak sites in 13 patients ( 7 ). All leakage sites were located in the visceral pleura; segment S4 was the most frequent site (8/13, 61.5%), followed by segments S3 and S6 (two cases each) and segment S5 (one case), and most lesions were bullae-related (11/13, 84.6%). Apart from this large case series, only a small number of case reports have directly documented actual pulmonary air-leak sites in catamenial pneumothorax. Reported locations include the right middle lobe, right upper lobe, and, less commonly, the left upper lobe, with several cases demonstrating pathological confirmation of ectopic endometrial tissue at the leak site ( 9 - 14 ). Our findings closely paralleled these observations because actual air leaks were also identified intraoperatively in 9 of 15 patients (60.0%), most often in the middle lobe and three-lobe confluence. The novelty of the present study does not lie in identifying segments S4 or S6 as common anatomical sites for pulmonary air leaks, as this distribution has been reported previously. Rather, the clinical significance of our findings lies in the high frequency of intraoperative identification of active pulmonary air-leak sites in consecutive patients with pathologically confirmed catamenial pneumothorax. In particular, pulmonary air leakage was identified in 9 of 15 patients and in 9 of 12 patients who underwent surgery while air leaks were confirmed or clinically suspected. Endometrial tissue was confirmed in all resected pulmonary specimens corresponding to the identified air-leak sites. These findings suggest that, when surgery is performed during active or clinically suspected ongoing air leakage, the responsible pulmonary lesions may be identified more frequently than previously appreciated in routine thoracic surgical practice. The potential clinical importance of identifying pulmonary air-leak sites is also supported by recurrence patterns. In our cohort, recurrence occurred in only 1 of 9 patients (11.1%) with an intraoperatively identified pulmonary air-leak site. Similarly, although not highlighted in the main text, the supplemental data of Tsuboshima et al. showed recurrence in only 1 of 13 patients (7.7%) with confirmed pulmonary air-leak sites ( 7 ). Although these observations are limited by the small sample size and retrospective study design, they suggest that surgery during active or clinically suspected ongoing air leakage may allow direct identification and removal of the responsible pulmonary lesion. This approach may help reduce the risk of recurrence when combined with appropriate coverage of the resected lung surface, treatment of diaphragmatic lesions, and postoperative hormonal suppression where indicated. Postoperative recurrence remains a major clinical concern in patients with catamenial pneumothorax. In the present case series, recurrence developed in 5 of the 15 patients (33.3%), and the 1- and 2-year recurrence-free survival rates were 85.6% and 63.4%, respectively. These findings are broadly consistent with those of previous reports showing that postoperative recurrence is frequent in patients with catamenial pneumothorax. For example, previous studies have reported postoperative recurrence rates ranging from 27% to 40% after surgical treatment for catamenial pneumothorax and related thoracic endometriosis during follow-up periods of approximately 3 years ( 2 , 15 , 16 ). More recently, Ochi et al. reported 1- and 2-year postoperative recurrence rates of 13.8% and 19.3%, respectively ( 6 ). The relatively higher recurrence rate in our cohort may partly reflect differences in surgical strategy. In their series, diaphragmatic lesions were generally resected or plicated, pulmonary lesions were resected when present, and the lung surface was covered with oxidized regenerated cellulose sheets to prevent pneumothorax recurrence and pleural adhesions ( 6 ). In contrast, the surgical strategy in our cohort was not standardized. Among the five patients who developed postoperative recurrence, four had undergone treatment of diaphragmatic lesions alone. In these cases, the responsible pulmonary lesions may not have been adequately inspected or resected, and the recurrent pneumothorax may have originated from an unrecognized pulmonary source. Conversely, one patient with recurrence underwent resection of an identified pulmonary air-leak lesion alone, without additional coverage of the staple line or treatment of diaphragmatic lesions. This case suggests that simple pulmonary wedge resection alone may be insufficient for recurrence prevention and that additional procedures, such as coverage of the resected lung surface and appropriate management of diaphragmatic lesions, may be necessary. Postoperative hormonal therapy may provide additional benefits, although the evidence remains inconsistent. Tsuboshima et al. demonstrated that postoperative hormonal therapy significantly reduced recurrence ( 7 ), whereas Kim et al. found a lower recurrence rate without statistical significance and identified diaphragmatic resection as a protective factor ( 17 ). In our cohort, postoperative hormonal therapy was administered to 12 patients, including all five who developed recurrence. Although hormonal therapy may suppress the activity or progression of endometriotic lesions, it may be inadequate to prevent the rupture of pre-existing bullae or blebs. Taken together, these findings suggest that the prevention of recurrence in catamenial pneumothorax may require careful intraoperative management of thoracic lesions, including pulmonary air-leak sites and diaphragmatic abnormalities, and postoperative hormonal suppression tailored to individual patients. A history of pelvic endometriosis was documented in 5 of 15 patients (33.3%) in our cohort. Although this proportion was slightly lower than those reported in previous surgical series, studies of catamenial pneumothorax and thoracic endometriosis-related pneumothorax have reported variable proportions of pelvic endometriosis, ranging from approximately 37% to 56%, reflecting differences in patient populations, diagnostic definitions, and the extent of gynecological evaluation ( 4 , 7 , 17 ). Because systematic gynecological screening was not performed in all patients in the present study, pelvic endometriosis may have been underestimated. This study has some limitations. First, this was a retrospective, single-center study with a small number of patients; therefore, the frequency and distribution of intraoperative air-leak sites may have been affected by institutional surgical practices and patient selection. Second, the surgical strategy was not standardized during the study period. Decisions regarding pulmonary wedge resection, reinforcement of pulmonary staple lines, treatment of diaphragmatic lesions, and postoperative hormonal therapy were made according to the intraoperative findings and the surgeon’s judgment. Accordingly, we could not determine which surgical procedure was the most effective in preventing recurrence. Third, although endometrial tissue was identified in the resected S1 and S6 bullous lesions in our cohort, the strict distinction between endometriosis-related bullous lesions and coincidental primary spontaneous pneumothorax lesions remains challenging. Because visceral pleural endometriotic lesions and pulmonary cystic lesions of primary spontaneous pneumothorax can overlap anatomically, particularly in segments S1 and S6, the possible coexistence of primary spontaneous pneumothorax should be considered when interpreting bullous lesions in catamenial pneumothorax ( 6 , 18 ). Fourth, pathological confirmation was not available for all macroscopically identified diaphragmatic lesions because diaphragmatic tissue was not obtained from some patients. Finally, because some patients underwent surgery after repeated episodes or prolonged air leakage, the ability to identify the true initial leakage site may have varied according to disease timing and the condition of the pulmonary lesion during surgery. Despite these limitations, our findings suggest that surgery performed during active or clinically suspected ongoing air leakage may increase the likelihood of identifying the site of a pulmonary air leak. Future multicenter prospective studies should evaluate whether catamenial pneumothorax can be diagnosed more objectively by combining macroscopic diaphragmatic abnormalities with direct intraoperative confirmation of pulmonary air leaks or pathological evidence of endometrial tissue in pulmonary lesions. Future studies should also determine the optimal management of diaphragmatic lesions and evaluate the effectiveness of additional recurrence-prevention strategies, such as pulmonary reinforcement, pleural coverage, and pleurodesis, particularly in patients for whom no responsible pulmonary air-leak site can be identified intraoperatively. This subgroup had the highest recurrence risk in our cohort and may require a more comprehensive surgical and postoperative management strategy. This approach may contribute to the development of evidence-based diagnostic and therapeutic strategies for catamenial pneumothorax.

This retrospective surgical series examined intraoperative pulmonary air-leak sites in 15 patients with pathologically confirmed catamenial pneumothorax. Diaphragmatic lesions were observed in all patients, whereas active pulmonary air leakage was identified in nine patients, most commonly in segment S4 and the three-lobe confluence. These findings suggest that pulmonary lesions may directly contribute to the development of pneumothorax in selected patients and that intraoperative assessment should not be limited to the diaphragm. In patients undergoing surgery during active or clinically suspected ongoing air leakage, careful inspection of the middle lobe and adjacent interlobar surfaces may help identify responsible pulmonary lesions and guide more complete surgical treatment. Further prospective multicenter studies are required to determine whether standardized evaluation and management of pulmonary and diaphragmatic lesions can reduce postoperative recurrence.

The article’s supplementary files as