Thoracic endometriosis is a rare disease responsible for catamenial pneumothorax. The immunohistochemical features of thoracic endometriosis are not well understood. An immunohistochemical examination of 84 diaphragmatic specimens of catamenial pneumothorax using antibodies against estrogen receptor (ER), progesterone receptor (PgR), CD10 and smooth muscle actin (SMA) was conducted. The endometrial tissue was small, and focally located around the chasm of the tendon on the side of the thoracic cavity. Endometrial stroma were detected in 84/84 (100%) of the specimens, endometrial glands were detected in 21/84 (25%) and smooth muscle was detected in 1/84 (1.2%). The endometrial stroma exhibited positive staining for ER in 74/84 (88.1%) of the specimens, PgR in 84/84 (100%), CD10 in 74/84 (88.1%) and SMA in 46/84 (54.8%). Because thoracic endometriosis is small in size, and only 25% of the resected tissue specimens were accompanied with the endometrial gland, an immunohistochemical analysis can be useful for their detection. The fact that over half of the thoracic endometrial stroma showed positive staining for SMA, and the existence of thoracic endometriosis accompanied by smooth muscle, indicated that some part of the thoracic endometriosis may have the ability to differentiate into smooth muscle, although further studies are needed to confirm this hypothesis.
Catamenial pneumothorax is one of the causes of spontaneous pneumothorax in females. Catamenial pneumothorax is thought to be a rare disease, reported to account for less than 1% of cases of spontaneous pneumothorax in female patients in the 1980s. Because physicians have paid more attention to this disease, 20–30% of females with spontaneous pneumothorax have recently been diagnosed with catamenial pneumothorax.[3-5] Most cases of catamenial pneumothorax involve spontaneous pneumothorax caused by thoracic endometriosis.[6, 7]
Although the number of patients with catamenial pneumothorax is increasing, the immunohistochemical features of thoracic endometriosis have not been fully investigated. Normal and ectopic endometrial gland and stroma have been reported to show positive staining for the estrogen receptor (ER), and progesterone receptor (PgR).[8, 9] Recent studies have shown that CD10 is a useful immunohistochemical marker of both normal and ectopic endometrial stroma.[10, 11] Furthermore, normal endometrial stroma and endometrial stromal sarcoma were reported to show positive staining for smooth muscle actin (SMA).
The aim of our study was to clarify the features of thoracic endometriosis using immunohistochemical staining.
Materials and Methods
Among the 570 female patients who underwent surgery for spontaneous pneumothorax at Nissan Tamagawa Hospital during the five years between January 2006 and December 2010, 84 (14.7%) patients diagnosed with catamenial pneumothorax were selected. The diagnosis of catamenial pneumothorax was made when all of the pneumothorax episodes occurred between 24 h before and 72 h after the initiation of menses. All of the operations were performed under video-assisted thoracoscopy. The diaphragm was partially resected using an endoscopic stapler and repaired with hand sutures. All patients were non-menopausal females. Their mean age at the time of surgery was 37.4 ± 5.4 years (24–50). A total of six (7.1%) patients had a history of tobacco consumption. The number of spontaneous pneumothoraxes prior to surgery was 3.8 ± 1.7 (1–10), and the period from the first occurrence of a spontaneous pneumothorax to surgery was 14.2 ± 13.9 months (0–72). Forty-eight patients (57.1%) were diagnosed with pelvic endometriosis. The mean serum CA125 level was 38.0 ± 36.5 (6–291) IU/mL. The operative findings revealed holes in the diaphragm in 71 (84.5%) patients and nodules on the diaphragm in 31 (36.9%) patients (Table 1). The study was approved by the institutional review board of Nissan Tamagawa Hospital.
Table 1. Characteristics of the study population
37.4 ± 5.4 years
The number of the pneumothorax until the operation
3.8 ± 1.7
The period from the first pneumothorax to operation
14.2 ± 13.9 months
Presence of pelvic endometriosis
The value of serum CA125, U/mL
38.0 ± 36.5
Holes in the diaphragm
Nodules in the diaphragm
The diaphragmatic tissue samples were fixed in 10% formalin and processed in a routine manner for light microscopy. The tissue sections obtained from paraffin-embedded tissue blocks were stained with hematoxylin-eosin.
Immunohistochemical staining was performed according to the Envision method using antibodies against ER, PgR, CD10, and SMA. After deparaffinization, 4-μm thick sections were treated with the high pH type target retrieval solution (Target Retrieval Solution pH 9, Dako, Glostrup, Denmark) for 40 min at 98°C. The sections were then washed with buffer solution, and the primary antibodies were applied for 30 min. The antibodies used consisted of an anti-ER antibody (clone ER1D5, dilution 1:50, Dako), anti-PgR antibody (clone PgR 636, dilution 1:800, Dako), anti-CD10 antibody (clone CD10 56C6, dilution 1:40, Dako) and anti-SMA antibody (clone SMA 1A4, dilution 1:100, Dako). After washing the sections with buffer solution, the dextran polymer reagent conjugated with peroxidase and secondary antibodies (EnVision, Dako) was applied for 30 min. The sections were incubated with diaminobenzidine for two minutes. After washing them in water, these samples were counterstained with Mayer's hematoxylin for two minutes.
Controls were obtained from the diaphragm of male autopsy cases (n = 20) and the diaphragm of female autopsy cases (n = 20). In both groups, an immunohistochemical examination using antibodies against the ER, PgR, CD10, and SMA were conducted in the same manner as described above.
The endometrial tissue was small, and was located around the chasm of the tendon of the diaphragm on the side of the interpleural surface. No endometrial tissue was detected on the interperitoneal surface. The endometrial tissue was located on the tendon in 31/84 (36.9%) specimens ( Fig. 1) and between the tendon in 53/84 (63.1%) specimens (Fig. 2). Endometrial stroma were detected in 84/84 (100%) of the specimens, endometrial glands were detected in 21/84 (25%), and smooth muscle was detected in 1/84 (1.2%) (Table 2).
Table 2. Summary of the histology of thoracic endometriosis
Number of cases (percentage)
83 (98.8%)/1 (1.2%)
On the tendon/Between the tendon
31 (36.9%)/53 (63.1%)
The basic histological findings of the cases were similar. The endometrial stroma consisted of round to oval stromal cells with round to oval nuclei and little cytoplasm. Little infiltration of inflammatory cells was observed around the endometrial stroma. Few changes were observed in the diaphragmatic tissue around the endometrial stroma except for hemorrhage in 19/84 (22.6%) of the specimens. The endometrial glands consisted of round to oval glands lined with pseudostratified columnar epithelium with elongated nuclei and eosinophilic cytoplasm. The endometrial glands were circumscribed by the endometrial stroma (Fig. 2). The smooth muscle consisted of small, stellate, and short fascicular collections of spindle cells with elongated nuclei and eosinophilic cytoplasm. The smooth muscle exhibited focal microscopic foci and was located successively with the endometrial stroma (Fig. 3).
The diaphragms of male autopsy cases exhibited negative staining for the ER in 20/20 (100%) of the specimens, PgR in 20/20 (100%), CD10 in 20/20 (100%), and SMA in 20/20 (100%). Similarly, the diaphragms of female autopsy cases exhibited negative staining for the ER in 20/20 (100%) of the specimens, PgR in 20/20 (100%), CD10 in 20/20 (100%), and SMA in 20/20 (100%).
The endometrial stroma exhibited positive staining for the ER in 74/84 (88.1%) of the specimens, PgR in 84/84 (100%), CD10 in 74/84 (88.1%), and SMA in 46/84 (54.8%). The endometrial glands exhibited positive staining for the ER in 18/21 (81.8%) of the specimens, PgR in 18/21 (81.8%), CD10 in zero and SMA in zero. The smooth muscle exhibited positive staining for the ER in 1/1 (100%) of the specimens, PgR in 1/1 (100%), CD10 in 1/1 (100%), and SMA in 1/1 (100%) (Table 3).
Table 3. Immunohistochemical findings of thoracic endometriosis
Endometrial stroma n = 84
Endometrial gland n = 21
Smooth muscle n = 1
Staining for estrogen receptor
Staining for progesterone receptor
Staining for CD10
Staining for smooth muscle actin
In the present study, over 80% of the thoracic endometrial glands and stroma showed positive staining for the ER and PgR. The thoracic endometrial stroma showed positive staining for CD10 in 74/84 (88.1%) of the specimens. Only 21/84 (25%) of the thoracic endometriosis specimens showed evidence of the endometrial gland. Approximately half (46/84; 54.8%) of the thoracic endometrial stroma showed positive staining for SMA. Smooth muscle was detected in 1/84 (1.2%) of the thoracic endometriosis samples.
Sumathi et al. recently showed that 23/25 (92%) specimens of ectopic endometrial stroma that did not include thoracic endometriosis exhibited positive staining for CD10. They concluded that the immunohistochemical staining for CD10 may be of value for confirming a diagnosis of ectopic endometriosis. In the present study, 74/84 (88.1%) of the thoracic endometrial stroma specimens exhibited positive staining for CD10. Immunohistochemical staining for CD10 may be of value for diagnosing thoracic endometriosis, similar to other types of ectopic endometriosis.
Only 21/84 (25%) of the cases of thoracic endometriosis were accompanied by endometrial gland. The reasons why we diagnosed stroma not accompanied by gland as endometriosis were as follows: (i) the stromal cells not accompanied by gland were very similar morphologically to the stromal cells accompanied by gland; and (ii) all of the cases examined in the present study were clinically diagnosed as catamenial pneumothorax, and a differential diagnosis of the stroma on the diaphragm is unlikely to be anything other than thoracic endometriosis.
The majority of cases of thoracic endometriosis are thought to be derived from pelvic endometriosis that has migrated through the peritoneal cavity. The ratio of thoracic endometriosis accompanied by pelvic endometriosis was reported to be 18–84%.[6, 13] This ratio may be underestimated, because a laparoscopic examination was not routinely performed to detect pelvic endometriosis in those reports.
The small ratio of the thoracic endometriosis that was accompanied by endometrial gland was speculated to be due to the following: (i) only endometrial stroma had migrated from pelvic endometriosis in many cases of thoracic endometriosis; and (ii) the endometrial gland is derived from the endometrial stroma, and the differentiation from endometrial stroma into endometrial gland is difficult in the thoracic cavity. Other potential explanations may exist, but are not suggested by the findings of the present study.
We found that over half of the thoracic endometrial stroma showed positive staining for SMA. Franquemont et al. showed that 10/10 (100%) samples of normal endometrial stroma and 10/12 (83.3%) samples of endometrial stromal sarcoma showed positive staining for SMA. They concluded that normal endometrial stroma and neoplastic endometrial stroma may have the ability to differentiate into smooth muscle. One study of five cases of thoracic endometriosis analyzed immunohistologically reported that all cases of endometrial stroma showed positive staining for SMA and 1/5 (20%) of the cases were accompanied by smooth muscle. Although a detailed discussion of these results was lacking, the results are consistent with the present findings. Since 1/84 (1.2%) cases of thoracic endometriosis was accompanied by smooth muscle in the present study, a small part of the thoracic endometrial stroma may have the ability to differentiate into smooth muscle.
Smooth muscle accompanied by the presence of endometrial tissue has been reported in cases of ectopic endometriosis involving types other than thoracic endometriosis. Smooth muscle is present in 88.2% of cases of deeply infiltrative endometriosis, 76% of cases of peritoneal endometriosis and 17.7% of cases of ovarian endometriosis. The origin of smooth muscle has been speculated. First, endometrial stroma has the ability of stem cells to differentiate into smooth muscle (many reports use the term ‘smooth muscle metaplasia’).[15, 18] Second, smooth muscle originates from the stroma surrounding the implanted endometrial tissue. The latter hypothesis is not supported in cases of thoracic endometriosis because there is no stroma on the tendon of the diaphragm.
In conclusion, this large-scale study to evaluate the immunohistochemical findings of thoracic endometriosis showed that immunohistochemistry is useful for histologically diagnosing thoracic endometriosis.
We thank Mitsugi Hideyuki (Department of Pathology, Nissan Tamagawa Hospital, Tokyo, Japan) for the skillful technical assistance.