SMARCA4-deficient dedifferentiated ovarian carcinoma with rapid progression during adjuvant platinum-based chemotherapy: a case report and implications for molecular-driven treatment strategies

Introduction

Dedifferentiated ovarian carcinomas (DDOC) are uncommon malignancies characterized by the presence of both undifferentiated and differentiated carcinoma components. These tumors exhibit aggressive biological behavior and are often diagnosed at later stages. ¹ Due to their rarity, limited data exist regarding their optimal management and response to conventional therapies. For patients diagnosed in the early stages, the standard of care does not differ from other aggressive ovarian carcinoma subtypes. It typically involves staging surgical procedures followed by adjuvant platinum-based chemotherapy (P-CHT). However, this treatment appears to be ineffective for this specific tumor subtype. ² A case report details the experience of a 55-year-old woman initially diagnosed with early-stage high-grade endometrioid carcinoma. Despite undergoing P-CHT, she exhibited rapid progression, leading to death 4 months after her diagnosis. Retrospective molecular analyses, which included the detection of pathogenic variants in the SMARCA4 gene, prompted a second expert pathological review that ultimately changed her diagnosis to DDOC. The presented case emphasizes the importance of integrating molecular analyses into diagnostic procedures. It highlights the urgent clinical need for more individualized treatment strategies reflecting the unique molecular landscape of this rare tumor. This case report conforms to CARE guidelines ³ (see Supplemental Material 1).

Case description

Clinical information

A 55-year-old woman was referred to our institution for abnormal vaginal bleeding and lower abdominal pain. An ultrasound examination revealed a unilocular-solid anechoic pelvic mass measuring 140 × 100 × 100 mm, with a thicker, vascularized wall, and a solid component measuring 28 × 22 × 25 mm that exhibited a color Doppler scale of 3. A CT scan confirmed significant enlargement of the left adnexa and indicated no distant abdominal or chest metastases (see Figure 1(a)). Based on the results of the perioperative histopathological examination, which indicated a malignant tumor likely representing serous carcinoma, the patient underwent complete staging surgical procedures, including pelvic and paraaortic lymphadenectomy. The final histopathological report confirmed the diagnosis of invasive high-grade endometrioid carcinoma of the left ovary, classified as stage IA according to the FIGO system. Due to the high-grade tumor, six cycles of adjuvant P-CHT were recommended and initiated after surgery. During the third cycle of P-CHT, an abdominal ultrasound was performed due to the worsening of her overall condition. This examination revealed multiple hyperechoic lesions in the liver as well as a solid, semi-fixed mass in the pelvis measuring 80 × 60 mm, located on the vaginal stump. CT scans confirmed the presence of multiple liver lesions ranging from a few millimeters to 40 mm, as well as various pelvic expansions, with the largest measuring 100 × 120 mm above the bladder (see Figure 1(b)). Histological examination of a pelvic biopsy confirmed infiltration by high-grade carcinoma identical to the initial histology. The patient’s overall condition rapidly deteriorated, ultimately resulting in her death 4 months after the initial diagnosis.

OPEN IN VIEWER

Figure 1.

Overview of coronal CT images of the abdomen and pelvis obtained after intravenous contrast administration in the portal venous phase. On the left (a), the staging examination reveals a large, predominantly cystic mass arising from the left adnexa, located superior to the urinary bladder. On the right (b), the follow-up examination during early recurrence of ovarian carcinoma demonstrates heterogeneously enhancing lesions in the pelvis, adjacent to the aortic bifurcation, and in the liver, consistent with multifocal metastatic spread.

Next-generation sequencing and immunohistochemical profiling results

As part of a research project focused on comprehensive molecular analyses of ovarian carcinomas with varying responses to P-CHT, a tumor sample of this case was retrospectively analyzed using next-generation sequencing (NGS) and immunohistochemistry (IHC). The NGS analysis of the undifferentiated component identified two likely oncogenic variants in the SMARCA4 gene, which are predicted to lead to a loss of normal protein function, specifically NM_001128849.3:c.2632_2633del (p.Met878AspfsTer66) and NM_001128849.3:c.3575G>A (p.Arg1192His). In addition, oncogenic or likely oncogenic variants were detected in the ARID1A, PTEN, PIK3CA, and CTCF genes. The undifferentiated component of the tumor exhibited microsatellite stability (1.64%) and a low tumor mutation burden (6.29 mutations/Mb), consistent with proficiency in the mismatch-repair (MMR) protein, as measured by IHC (see Table 1 for details). Further immunohistochemical analysis of the undifferentiated component assessed the expression of programmed cell death ligand 1 and revealed the value of a combined positive score 30 (CPS PD-L1). The density of tumor-infiltrating lymphocytes (TILs) was high in both intraepithelial and stromal tumor components. However, the expression of Schlafen-11 (SLFN-11), measured immunohistochemically by H-score, showed entirely negative results (see Table 2 for details).

OPEN IN VIEWER

Table 1.

Summary of the key results of the DDOC somatic analyses: results of the next-generation sequencing (undifferentiated component).

Gene	Transcript	HGVSc	HGVSp	VAF (%)
ARID1A	NM_006015.6	c.3156T>G	p.Tyr1052Ter	32.3
ARID1A	NM_006015.6	c.2989-7_2994del	p.?	19.9
SMARCA4	NM_001128849.3	c.2632_2633del	p.Met878AspfsTer66	38.2
SMARCA4	NM_001128849.3	c.3575G>A	p.Arg1192His	32.7
PIK3CA	NM_006218.4	c.3140A>G	p.His1047Arg	31.8
PTEN	NM_000314.8	c.161dup	p.Arg55LysfsTer8	29.8
PTEN	NM_000314.8	c.377C>A	p.Ala126Asp	32.9
CTCF	NM_006565.4	c.1342C>T	p.Arg448Ter	32.5

DDOC, dedifferentiated ovarian carcinomas; HGVS, human genome variation society; VAF, variant allele frequency.

OPEN IN VIEWER

Table 2.

Summary of the key results of the DDOC somatic analyses: results of immunohistochemical analyses.

Marker	LG endometrioid adenocarcinoma (90%)	Undifferentiated component (10%)
p53	No aberrant expression (“wild-type” pattern)	No aberrant expression (“wild-type” pattern)
WT-1	Negative	Negative
PAX-8	Positive	Negative
ER	100%	0%
PR	80%	0%
Ki-67	30%	50%
MMR	NA	Proficient
PD-L1 (CPS)	NA	30
SLFN-11	NA	Negative (H-score 0)
TILs (No/mm2)		Intratumorally	Stromal
CD3+	NA	401,055409	3270,27785
CD8+	NA	352,6993099	2080,13597
CD4+	NA	226,920686	1493,23684
CD164+	NA	771,1132071	2079,84286
Morphology a	NA	10	25

a

The average percentage of area infiltrated by mononuclear inflammatory cells is relative to the intratumoral intraepithelial and stromal areas, respectively.

DDOC, dedifferentiated ovarian carcinomas; ER, estrogen receptors; LG, low-grade; MMR, mismatch repair system proteins; No/mm², number per square millimeter; PD-L1 (CPS), combined positive score for programmed cell death ligand 1; PR, progesterone receptors; SLFN11, Schlafen 11; TILs, tumor-infiltrating lymphocytes; NA, not applicable.

Second expert pathological review

Based on the results of retrospective NGS analyses, the expert pathological review revised the initial diagnosis to dedifferentiated carcinoma of the ovary, which included a minor undifferentiated component that constituted 10% of the tumor. This undifferentiated component showed a loss of PAX8 expression and exhibited negative expression for both estrogen (ER) and progesterone receptors (PR). By contrast, the major differentiated component of the tumor, comprising 90%, consisted of endometrioid adenocarcinoma structures that displayed squamous differentiation. This component preserved the expression of ER, PR, and PAX8, and presented a p53 wild-type pattern with patchy p16 expression and WT1-negative status (see Figure 2).

OPEN IN VIEWER

Figure 2.

Representative histopathological images of dedifferentiated ovarian carcinoma (100× magnification). The representative images of histopathological findings are demonstrated (100× magnification). The tumor contained differentiated (a) and undifferentiated (b) components. The undifferentiated component showed a negative Schlafen-11 expression (c), the complete loss of estrogen receptors (d), the combined positive score for PD-L1 value of 30 (e), and a high density of CD3+ tumor-infiltrating lymphocytes (f).

Discussion

Dedifferentiated carcinomas generally have a poor prognosis due to their aggressive nature, regardless of their site of origin. In the female genital tract, these tumors are well described in the endometrium (dedifferentiated endometrial carcinoma, DDEC), where they were first reported in 2006. ⁴ However, data regarding clinical manifestations and treatment outcomes of DDOC remain scarce, mainly due to their rarity and potential for misdiagnosis.

Establishing an accurate histopathological diagnosis is not always straightforward. Specifically, distinguishing between the undifferentiated and poorly differentiated components of endometrioid adenocarcinoma can be challenging in some cases. Consequently, DDOC is most commonly misdiagnosed as high-grade endometrioid carcinoma, although high-grade serous carcinomas or neuroendocrine tumors may also appear in the differential diagnosis.^5,6 The undifferentiated components, usually responsible for aggressive biological behavior irrespective of their size,^2,7 frequently exhibit specific molecular features. These characteristics include loss of function in SWI/SNF complex units, particularly based on SMARCA4, SMARCB1, or concurrent ARID1A/ARID1B inactivation.^6,8,9 However, markers indicating a potential SWI/SNF complex function loss might not be routinely examined. In our case, the discovery of two pathogenic variants in the SMARCA4 gene resulted in a retrospective diagnosis change to DDOC.

Fulminant progression during adjuvant P-CHT, as observed in our case, is exceptionally rare. The recommendation of adjuvant P-CHT—regardless of a FIGO stage IA and after the complete surgical staging—was based on treatment guidelines for high-grade epithelial ovarian carcinoma aimed at reducing the risk of recurrence.^10,11 It remains questionable whether knowing the correct diagnosis of DDOC would alter the decision regarding adjuvant therapy. Since DDOC is often diagnosed at advanced stages, data on the outcomes of the adjuvant treatment in early stages are limited. To our knowledge, fewer than 10 DDOC cases diagnosed in FIGO stage I, with available treatment and follow-up data, have been reported. In the largest series of 23 DDOC patients, 5 were diagnosed with FIGO stage I, and 4 succumbed to recurrent disease within 2 years. Similar to our case, all patients who received adjuvant chemotherapy experienced disease progression during chemotherapy, with two dying within 4 months. ² Thus, despite their high growth potential, the standard P-CHT appears ineffective in DDOC.

In our case, the observed resistance to P-CHT correlated with completely negative IHC results for SLFN-11 expression in the undifferentiated component (see Table 2 and Figure 2(d) for details). The nuclear protein SLFN11 has been recently discovered as the causal and dominant genomic determinant of response to replication stress caused, for example, by anticancer agents damaging DNA, including platinum compounds. ¹² The absence of the SLFN-11 expression has been associated with resistance to DNA-damaging agents, making it a promising predictive biomarker. ¹³ Although SLFN-11 and the SWI/SNF complex, both deficient in our case, play important roles in regulating DNA damage response and transcription, a direct link between SLFN11 and the SWI/SNF complex has not yet been described or unequivocally established.

Given the P-CHT’s inefficiency, the strategy for the systemic treatment of DDOC requires modifications. One promising approach involves incorporating immunotherapy through checkpoint inhibitors (CPi), based on the biological similarities with DDECs, which exhibit MMR deficiency in up to two-thirds of cases. ¹⁴ Despite strong evidence suggesting that MMR deficiency serves as a predictive marker for CPi across various solid tumors, other factors can also indicate a positive response to immunotherapy, including high expression of PD-L1—a known immunosuppressive checkpoint protein—and intense T-cell infiltration. ¹⁵ Although DDOCs appear to be less frequently MMR-deficient than DDECs (6 out of 23 cases in Tessier et al.), additional markers may suggest a beneficial role for CPi in treatment strategies. As shown in Table 2, IHC results in our case revealed not only MMR proficiency, but also positive PD-L1 expression measured by combined positive score (CPS-PDL1), alongside a high proportion of both intratumoral and stromal TILs in the undifferentiated component (see Figure 2(e) and (f)). These findings support the assumption that incorporating CPi into systemic treatment may enhance efficacy. Another approach to improve treatment outcomes is to target the altered SWI/SNF complex through developing molecular targeted therapies. ¹⁶

Given the rarity of DDOC, expecting data from large randomized prospective clinical trials to alter established treatment standards is unrealistic. Therefore, a precision oncology treatment approach—developing an individual treatment plan based on detailed tumor analyses searching for alterations with already existing targeting treatment—seems most appropriate. In addition to IHC results, the NGS analysis revealed several potentially targetable alterations. Inactivating variants in SMARCA4 and ARID1A, both core components of distinct SWI/SNF chromatin remodeling complex subunits, suggest epigenetic dysregulation that could be therapeutically utilized using EZH2 inhibitors. EZH2 inhibitors have already been approved for treating epithelioid sarcoma, which is characterized by inactivation of SMARCB1, another member of the SWI/SNF complex. Based on a similar rationale of exploiting the synthetic lethal interaction between SWI/SNF deficiency and PRC2 complex activity, several clinical trials are currently evaluating EZH2 inhibition in tumors with ARID1A or SMARCA4 loss. Notably, the GY014 phase II trial (NCT03348631) is evaluating tazemetostat in patients with recurrent ovarian or endometrial cancer harboring ARID1A alterations. In parallel, variants identified in PIK3CA and PTEN suggest activation of the PI3K/Akt/mTOR pathway, which might be targeted with either PI3K or mTOR inhibitors. Recently published data demonstrated that the PI3Kα-specific inhibitor alpelisib achieved an objective response rate of 28% and a disease control rate of 61% in patients with PIK3CA-mutated recurrent gynecologic cancers, including ovarian carcinoma, supporting biomarker-driven use of PI3K blockade. ¹⁷ Altogether, the molecular data not only refined the tumor classification but could also offer a biologically informed basis for exploring non-standard, targeted therapeutic strategies in this rare and aggressive entity in an upfront testing setting.

Conclusion

We present a rare case of SMARCA4-deficient DDOC that was initially diagnosed as high-grade endometrioid carcinoma. This case followed a highly unfavorable clinical course, with fulminant progression during adjuvant P-CHT, ultimately resulting in death 4 months after initial diagnosis, despite the disease being detected at an early stage. Retrospective molecular profiling allowed us to refine the initial morphological diagnosis and underscored the clinical need for more effective systemic treatment for this rare malignancy. NGS analysis revealed the SWI/SNF complex alteration as well as PI3K/Akt/mTOR pathway activation, both of which are linked to already existing therapeutic options. Immunohistochemical analysis demonstrated a loss of SLFN-11 expression in the undifferentiated component, correlating with the real-life resistance to P-CHT. Conversely, positive PDL1-CPS values and a high density of TILs support the potential effectiveness of immunotherapy using CPi. Future research should focus on elucidating the molecular mechanisms underlying P-based chemotherapy resistance in patients with this formidable diagnosis. In addition, establishing new systemic treatment strategies that integrate immunotherapy or molecularly targeted treatments, with or without chemotherapy, is essential. In our case, a precise oncologic approach consisting of detailed somatic tumor profiling would have facilitated the development of a more individualized treatment plan had it been employed at the time of diagnosis.

Supplemental Material