Thoracic Gorham-Stout disease masquerading as an ovarian malignancy: a differential diagnosis in non-malignant causes of raised Ca-125 and overview of this rare pathology

Abstract

Gorham-Stout disease (GSD) is thought to be due to uncontrolled proliferation of vascular and lymphatic structures within bone tissue causing destruction and osteolysis of bone. We present a patient in her mid-40s who reported chronic shoulder pain, a pleural effusion and irregular periods. Investigations showed osteolysis of her ribs, pleural effusions, an ovarian mass and a raised carbohydrate antigen 125 (Ca-125). She was subsequently diagnosed with GSD, and referred to gynaecology-oncology in consideration of potential ovarian malignancy. GSD is a diagnosis of exclusion that requires a high degree of clinical suspicion, as well as multiple investigations to achieve diagnosis. Clinicians rely on a small number of case reports to provide guidance for this. Therefore, this report provides an overview of a rare pathology, considers the differentials of a raised Ca-125 and describes how a pleural effusion, which links them both, alarmed us regarding an incidental finding of an ovarian cyst.

Background

Gorham-Stout disease (GSD) is a rare syndrome of bone osteolysis that was first described by Jackson in 1838 and defined by Gorham and Stout in 1955.^{1 2} The exact aetiology is unknown; however, the uncontrolled proliferation of vascular and lymphatic structures within bone tissue causing destruction and osteolysis of bone is the most widely accepted theory regarding the disease process. GSD most commonly involves the skull, shoulder, jaw, ribs, spine and pelvic girdle.³ Primary involvement of the spine has been described in approximately 50 cases.⁴ GSD can present with spontaneous fractures, localised swelling and most commonly—gradually increasing pain with deformity, as well as symptoms related to the location and anatomy of the lesion. For example, thoracic involvement has been associated with pleural effusions, chylothorax and therefore respiratory compromise, as in the case of our case.⁵ GSD is a diagnosis of exclusion that requires a high degree of clinical suspicion and is based on histological, radiological and clinical findings. In 1983, Heffez et al published eight criteria for a definitive diagnosis including biopsy with the presence of angiomatous tissue, absence of cellular atypical, minimal osteoblastic response or dystrophic calcification, evidence of local bone progressive osseous resorption, non-expansile non-ulcerative lesions, no involvement of viscera, osteolytic radiographic pattern and finally, negative hereditary, metabolic, neoplastic, immunological or infectious aetiology.⁶ It has been noted that the disease can stabilise or go into remission after a number of years. Complications include severe deformity, functional disability and pain.³ Prognosis is unpredictable, and poorer outcomes are associated with spine and skull base involvement causing neurological compromise. Treatments are guided by case reports and are aimed at symptom management. They include interferons, bisphosphonates, surgical resections, radiotherapy and sirolimus. Carbohydrate antigen 125 (Ca-125) is a cell surface antigen expressed in tissues such as the endometrium, peritoneum, pleura and pericardium. It is synthesised in response to mechanical stress and inflammation and may play a role in cell-mediated immune response. It is used as a survival predictor and monitoring tool in ovarian cancer. Due to the paucity of cases and formal guidelines, clinicians are guided by case literature. Therefore, documenting all presentations of GSD and response to therapy contributes to our knowledge of the disease while highlighting its importance as a differential in any patient presenting with osteolysis of bone as well as a raised Ca-125.

Case presentation

Our patient is a woman in her mid-40s who presented to the emergency department with worsening right-sided shoulder and pleuritic chest pain. She has no medical history, and her symptoms first began 5 years prior. She reported recurrent episodes of pleuritic chest, right shoulder and back pain lasting for weeks at a time. There was no clear trigger, and episodes occurred at random. She has no family history of bone disorders, malignancy or tuberculosis. Previous investigations included an MRI of the spine, dual energy X-ray absorptiometry scan, shoulder joint ultrasound and nerve conduction studies—all of which were unremarkable. Due to the pleuritic nature of her chest pain, initial investigations in the emergency department were carried out. This included a COVID-19 PCR, which was negative, and a D-dimer, which was raised at 12 035 (0–500 µg/L). Examination findings suggested C5–C8 cervical nerve root pain with no neurological weakness. Observations, ECG and baseline blood tests were otherwise unremarkable. She underwent chest X-ray in consideration of an apical lung mass causing neuropathic pain, which subsequently reported a hypoplastic right upper lobe, right-sided rib fractures, partial collapse of the right lower lobe, small bilateral pleural effusions and poor visualisation of the right second–fourth ribs (figure 1).

Figure 1.

Admission chest X-ray report: there is abnormal morphology of the right upper hemithorax. The right upper lobe appears relatively hypoplastic and the right second–fourth ribs are poorly visualised (resected). There are right-sided rib fractures including an acute/subacute fracture of the right fifth posterior rib and a healed fracture of the sixth rib. There is partial collapse of the right lower lobe with associated inferior displacement of the right hilum. There is a small right pleural effusion and small left pleural effusion, but no pneumothorax.

The patient underwent CT pulmonary angiogram (CTPA)due to the raised D-dimer and to detail the abnormal anatomical findings further. This was negative for pulmonary embolism (PE) or evidence of malignancy (figure 2). At the time, appearances were queried to be congenital. Due to this, further assessment of bone metabolism was performed. Results of note included a positive antinuclear antibody (titre: 80; pattern: speckled), raised IgM 2.82 (0.5–2.5 g/L), parathyroid hormone 9.5 (1.6–7.2 pmol/L) and CTx level 0.65 (0.1–0.5 µg/L)—a marker of bone turnover warranting referral to the rheumatology team. Although these findings were deemed to be significant, the breadth of abnormal findings posed quite a challenge with regard to immediate investigation requests. With the potential risk of thoracic malignancy taking initial precedence, the patient was urgently referred to the respiratory team, who considered a wide differential for her effusions including inflammatory serositis, catamenial infection and malignancy. Pleural aspiration drained 850 mL of blood-stained pleural fluid. Cytological analysis showed this to be a non-malignant exudate.

Figure 2.

CT pulmonary angiogram report: no pulmonary emboli, enlarged hilar or mediastinal lymph nodes. The tracheobronchial tree is patent. No significant lung nodules or masses present. There is a large right pleural effusion with passive atelectasis of the majority of the right lower lobe. There is no pneumothorax. The left pleural space is clear. The right ribs 2–5 are absent. There is discontinuation of the first rib. There is an old fracture/pseudarthrosis of the right posteromedial sixth rib. There is no destructive bone lesion.

A repeat MRI of the spine showed diffuse increased T1 and T2 marrow signal seen through C7, T1, T2 and T3 (figure 3). There was no nerve root impingement, and the spinal multidisciplinary team (MDT) felt that there was no structural spinal cause of her symptoms, additionally excluding Poland syndrome. Thoracoscopy was performed to investigate chylothorax, a common presentation of GSD. This demonstrated a small area of focal lymphangitis from which biopsies were taken; however, these did not show chylothorax, malignancy or angiomatous proliferation and therefore the diagnosis remained unconfirmed.

Figure 3.

MRI of the cervical spine report: there is fatty signal change within the C7–T3 vertebral bodies with partial sparing of their spinous processes.

Concurrent to the presentation and investigations already discussed, the patient additionally reported to her general practitioner a 4-month history of abdominal bloating, weight gain and irregular menstrual bleeding with occasional vasomotor symptoms. Her cervical smears are up to date and she has no personal or family history of breast or gynaecological malignancies. She is multiparous having undergone two previous caesarean sections. Transvaginal ultrasound was requested to examine for ascites, ovarian mass or Meig’s syndrome. This was reported as a right ovarian cyst, of possible haemorrhagic or endometriotic origin, with a trace of pelvic fluid (figure 4). Routine blood tests in the community revealed a raised Ca-125 at 146 U/mL. The patient was therefore urgently referred to the gynaecological-oncology team. CT of the abdomen and pelvis was requested to investigate for pelvic malignancy, which demonstrated a likely physiological right ovarian cyst measuring 2.8 cm. There were no abnormalities noted on clinical examination and the patient was referred for an extended panel of tumour markers, outpatient hysteroscopy and MRI of the pelvis to further evaluate the small cyst. Repeat Ca-125 was further raised at 181 U/mL with carcinoembryonic antigen (CEA), Ca 19-9 and germ cell markers within normal limits. Hysteroscopy was noted to be unremarkable with a benign endometrial biopsy. The MRI of the pelvis confirmed a simple 21 mm right ovarian cysts with no further gynaecological cause found for the raised Ca-125 (figure 5). The patient was discussed at the gynaecological-oncology MDT, concluding that there was no suspicion of a gynaecological malignancy and the raised Ca-125 was likely secondary to the known pleural effusion and erosive changes in the chest wall.

Figure 4.

Ultrasound of the pelvis report: right ovarian cyst, of possible haemorrhagic or endometriotic origin, with a trace of pelvic free fluid.

Figure 5.

MRI of the pelvis report: 21 mm simple, right ovarian cyst. There is no endometriosis. Raised carbohydrate antigen 125 may be secondary to the pleural effusions and erosive changes in the chest wall.

On further review of the patient’s historical imaging, previous MRI had revealed all ribs to be present, which leads the team to explore a destructive process responsible for the presentation. Opinion from a musculoskeletal radiologist was sought, and then came the first suggestion of GSD. Thoracoscopy and histopathology of the biopsied pleura reported non-specific acute and chronic inflammation with no evidence of malignant cells. D2-40, CD34, CD31 immunohistochemistry staining and Van Gieson staining of this biopsy showed no abnormal vessels or angiomatous proliferation. Standard and tuberculosis cultures were also negative. The patient then underwent a CT-guided biopsy of the C7 spinous process in an effort to obtain a tissue diagnosis, as suggested by the Lymphangiomatosis and Gorham Disease Alliance. This showed fibrous tissues with variably sized vessels partly lined by D2-40-positive endothelial cells—confirming the diagnosis of GSD (figures 6 and 7). The patient went on to report further deterioration in her symptoms, with breathlessness on lying flat and exertion likely due to recurrence of pleural effusion. Her ability to perform activities of daily living was now compromised and treatment was commenced.

Figure 6.

On the right side, small residual bone trabeculae are present with only partial osteoblast rimming. No oteoclast activity is seen. Other areas show replacement of bone by fibrous/fibrovascular tissue, which is pathognomonic of Gorham-Stout disease (magnification ×100).

Figure 7.

Within the fibrous tissue, large numbers of lymphatic vessels are seen highlighted by D2-40 immunohistochemical reaction, confirming the diagnosis of Gorham-Stout disease (magnification ×100).

Differential diagnosis

Our patient initially presented with intermittent cervical radiculopathy. Differentials for this include acute disc prolapse, torticollis, trauma, arthritis, malignancy and now GSD. Specifically, after MRI of the cervical spine, differentials of hyperparathyroidism, myelofibrosis and myeloma were considered. However, renal function, electrophoresis and histology were unremarkable. The absence of ribs can be congenital, such as in Poland syndrome or part of an osteolytic process, as seen in GSD. Differentials of exudative pleural effusions that were investigated include infection (parapneumonic, mycobacterial and fungal), PE, malignancy, connective tissue disease, Meig’s syndrome, chylothorax, sarcoidosis and finally, GSD. B-type natriuretic peptide (BNP) was also assessed to consider cardiac causes of pleural effusion. PE was strongly considered as a differential for pleuritic chest pain, shortness of breath and a pleural effusion. D-dimer was not performed on admission; however, this was later found to be considerably raised, highlighting another similarity between GSD and PE. Another important differential of both our patient’s symptoms and PE at the time includes COVID-19 infection. Ovarian carcinoma is an important differential for a patient presenting with irregular menstrual cycles, abdominal bloating, pleural effusion, a raised Ca-125 and an ovarian mass. In this case, the raised tumour marker was a red herring, which has a wide differential diagnoses including ovarian carcinoma, fibroids, endometriosis, pelvic infection, benign ovarian cysts, ascites, liver disease and pleural effusions.

Treatment

Following expert consultation, our patient was commenced on dual therapy with sirolimus (titrated to 3 mg once a day) and zoledronic acid infusions (4 mg 4 weekly) with a view to continue treatment for a minimum of 12 months. Sirolimus is a mechanistic target of rapamycin (mTOR) inhibitor that targets lymphangiogenesis in order to achieve disease control by reducing osteoclastic proliferation and preventing further bone destruction. Sirolimus requires vigilant monitoring for side effects, by way of sirolimus levels and routine blood tests. General side effects include abdominal pain, arthralgia, constipation, diarrhoea, fever, headache and tachycardia. More specific side effects include diabetes mellitus, dyslipidaemia, thrombosis and embolism, neoplasms, osteonecrosis, ovarian cyst, pancreatitis, pericardial effusion, respiratory disorders and increased risk of infection. Finally, side effects to monitor for include anaemia, electrolyte imbalance, haemolytic uraemic syndrome, leucopenia, neutropenia, thrombocytopenia and proteinuria. Due to the risk of immunosuppression, the patient also received the COVID-19 vaccine and Pneumocystis pneumonia prophylaxis.

Outcome and follow-up

Initially, the patient’s symptoms stabilised, and thus far, she has responded well to the zoledronic acid infusions and sirolimus. Interval thoracic ultrasounds have demonstrated a reduction in the pleural effusions, and the patient subjectively reports a significant improvement in her chest wall pain. She reported some tenderness over the upper cervical spine and a repeat MRI of the spine suggested possible progressive changes of C1, C2 and T2 with stable disease elsewhere and no cord or cauda equina compression. She also reported ongoing hot flushes, night sweats and amenorrhoea. Blood tests in primary care showed a raised follicle-stimulating hormone. The respiratory team felt that her menstrual abnormalities were not associated with sirolimus. Repeat pelvic ultrasound showed stable appearances of the uterus and right ovarian simple cyst. She was seen in the menopause clinic, by which point her menstrual cycle had resumed and therefore they will continue to monitor her symptoms with a consideration for hormone replacement therapy in the future if indicated. Following 12 months of treatment, lung function tests, routine blood tests, blood gas analysis and pelvic ultrasound have all been stable. Moving forward, the patient will require ongoing monitoring for medication side effects and disease progression, with recourse for MDT discussion in the event of ventilatory failure requiring spinal surgery.

Discussion

GSD, also known as vanishing bone disease, is a rare skeletal disorder of uncertain aetiology. The condition was first reported in 1838, when Jackson documented the disappearance of the humerus in a young man in ‘a boneless arm’.¹ Following his death, Jackson presented his dissection findings, confirming the osteolytic process that had occurred. In 1955, Gorham and Stout reviewed 24 cases, including 8 histopathological studies in which they found vascular anomalies occupying spaces within the bone. They proposed the hypothesis that angiomatosis of blood and lymphatic vessels was the cause of the progressive osteolysis.² The syndrome was then coined ‘Gorham-Stout disease’, otherwise known as vanishing bone disease, phantom bone disease or disappearing bone disease. Further studies accepted and embarked on this theory, most notably by Moller et al who reported that focally increased bone resorption was due to an increased number of hyperactive osteoclasts.³ Fewer than 300 cases have been reported.⁷ The exact mechanism driving osteoclast activity, lymphangiogenesis and osteolysis is still unclear. Speculations include hyperaemia, changes in local pH, interleukin-6 and osteoclast precursors, with increased sensitivity to humoral factors that secrete angiogenic factors in a paracrine and autocrine manner.^{7 8} Immunohistochemical markers reported in proliferating lymphatic endothelial cells in GSD include CD31, CD34, LYVE-1, D2-40 and VEGF.⁹

Presentation of GSD includes spontaneous fractures, swelling, pain and deformity of the skull, shoulder, jaw, ribs, spine and pelvic girdle. Investigations include blood tests, lung function studies, plain X-rays, ultrasound, bone scans, CT scans and MRI scans but ultimately, diagnosis is obtained with biopsy of the affected tissues revealing abnormal lymphatic tissue and characteristic bony changes.

The use of interferons, bisphosphonates, surgical resections, radiotherapy and sirolimus has been reported in the few cases of GSD documented within the literature. Interferon alpha is an immunomodulatory anti-angiogenic medical treatment that has been reported to result in disease regression in some case reports of GSD.⁸ The additional use of bisphosphonates is well established due to its anti-osteoclastic and pro-osteogenic effect; however, the sample population is scarce in size. Another treatment option is sirolimus—an mTOR inhibitor used in a variety of vascular malformations. While it inhibits signalling pathways specific to lymphatic growth, side effects are often unacceptable due to the wide activity of mTOR in normal physiology. Sirolimus has been noted to be a successful treatment of GSD as well as other lymphangiomatotic disorders. While there have only been approximately 24 reported cases of GSD treated with sirolimus, this has been shown to stabilise or reduce signs and symptoms of disease as well as improve functioning. Due to the rarity of GSD, lack of reliable evidence base and unpredictable nature of the treatment response, all cases should be shared, and specialist input should be sought to allow for multistep targeted therapy in a field where guidelines are absent. Chylothorax is a well-documented complication, which carries a poorer prognosis and may require pleurectomy and pleurodesis.^{8 9} Spinal lesions may be managed with radiotherapy, vertebroplasty or osteotomy as well as other surgical interventions.¹⁰ Although it is reported as benign, the prognosis of GSD is unpredictable for many reasons. First, there is variability as to whether lesions remain stable or if bone resorption eventually stops. Second, pulmonary and spinal involvement confer poorer prognosis when compared with those with limb or pelvic involvement.

There is no documented literature that suggests any corelation between GSD and abnormal tumour markers, more specifically Ca-125. Ca-125 is a tumour antigen used to exclude and monitor ovarian cancer. The raised Ca-125 was a red herring in this case, and combined with complex symptomatology, led to our patient undergoing an ultrasound, CT, MRI and hysteroscopy with endometrial biopsy to detail the significance of her simple ovarian cyst. Our patient presented with a history of abdominal bloating, irregular menstruation and weight gain. The Ca-125 was checked and raised, warranting further investigations to rule out an ovarian carcinoma. Initial imaging went on to reveal a subsequently benign ovarian mass, which nevertheless became an undue cause of concern. While other teams were confronted with multiple negative investigation findings, the pursual of a gynaecological-oncology primary pathology continued. Throughout this patient’s journey, ovarian cyst, ovarian fibroma, ovarian carcinoma and benign or malignant pleural effusion were all considered in the differential diagnosis.

Ca-125 is an antigen used in the diagnosis of ovarian cancers.¹¹ Ca-125 was first identified in 1981 as the target of 125th antibody produced against the ovarian cancer cell line.¹¹ Ca-125 is expressed in the endometrium, peritoneum, pleura and pericardium. Ca-125 levels can increase in both physiological and pathological conditions such as the menstrual cycle, first trimester pregnancy, ovarian carcinoma, fibroids, endometriosis, pelvic inflammatory disease, benign ovarian cysts, hepatitis, ascites, pleural effusions and lymphangioleiomyomatosis (LAM).^{11 12} It has been suggested that Ca-125 is synthesised in the mesothelial cells as a stress response to either fluid congestion (mechanical stress) or cell mediators (inflammatory stress).¹³ Howe et al comment on the importance of understanding how these biological mechanisms correlate with a wide range of conditions, for clinicians who are assessing the significance of elevated Ca-125 levels.¹³ Our case highlights the importance of appreciating and communicating to the patient, the limitations, implications and potential harmful consequences of using this particular ‘gold standard’ tumour marker. Given this, the authors acknowledge the challenges faced when using a non-specific tool to aid diagnosis. The Royal College of Obstetricians and Gynaecologists advise that Ca-125 is unreliable in differentiating benign from malignant masses in premenopausal women due to increased rates of false positives and reduced specificity.¹⁴ Usually, a serum Ca-125 assay is not required when a clear ultrasonographic diagnosis of a simple ovarian cyst has been made. They suggest that serial monitoring may be beneficial to assess for rapidly rising levels, and to discuss with a gynaecological oncology specialist if the assay is more than 200 U/mL.¹⁴ Due to unusual and complex circumstances, and an unclear initial ultrasound report, this tumour marker was requested and became somewhat of a red herring.

Ca-125 is known to be raised in LAM, a low-grade neoplasm due to proliferation of atypical smooth muscle cells. LAM is primarily found in women and features cystic lung destruction, chylothorax, respiratory failure and lymphatic abnormalities. LAM can be sporadic or associated with tuberous sclerosis, and has many clinical and pathological features that overlap with GSD. In a study of 241 female patients with LAM, almost one-quarter were found to have at least one elevated Ca-125. Higher levels were associated with pleural effusion and subsequently decreased with sirolimus therapy. The study concluded that LAM cells express Ca-125 and therefore we propose that GSD should be considered a differential for a raised Ca-125.¹² We document this correlation to bring to the attention of our clinical colleagues in an effort to raise awareness and assist in avoidance of unnecessary potential expensive and invasive tests. While investigating an unfamiliar disorder like GSD, it becomes challenging to know which pathological process might attribute to this raised tumour marker. MDTs were involved in order to determine the significance of the tumour marker and imaging findings. In conclusion, GSD should be considered as a differential for osteolysis. Further studies are required to assess whether GSD, like LAM, can be considered as a cause for a raised Ca-125.

Learning points.

• Gorham-Stout disease (GSD) is a rare osteolytic disorder that requires a high degree of clinical suspicion, multiple imaging modalities and a multidisciplinary approach in order to be diagnosed.

• Differential diagnoses of a raised carbohydrate antigen 125 (Ca-125) include benign and malignant ovarian masses, endometriosis, adenomyosis, pelvic infections, liver cirrhosis, ascites and pleural effusion. Therefore, a raised Ca-125 is not specific to ovarian carcinoma.

• Ca-125 can be raised in lymphangioleiomyomatosis. Further assessment is required to assess whether this non-specific marker can be attributed to a rare pathology such as GSD.

Ethics statements

Patient consent for publication

Obtained.