Paclitaxel and Carboplatin for Uterine Carcinosarcoma: A Path to Inclusion

Uterine carcinosarcoma (UCS), a rare and highly aggressive histologic subtype, which affects disproportionately Black women, represents < 5% of all uterine cancers, but stage for stage has a fatality rate substantially higher than endometrial cancer.¹ UCS is a biphasic tumor containing both epithelial and sarcomatous components, had been historically classified as a subtype of uterine sarcoma, and, as such, is treated with sarcoma regimens and excluded from trials targeting the broader uterine cancer patient population.² Although cellular lineage tracing studies are yet to be completed, recent genomic evidence suggests that UCS likely originates from a carcinoma precursor undergoing late sarcomatous dedifferentiation, rather than from two independent progenitors, epithelial and mesenchymal, as previously speculated.^2-4 In parallel with our developing understanding of this rare tumor's biology, treatment protocols are evolving, shifting from ifosfamide-based regimens with known activity in sarcoma to those more commonly used in the management of uterine (and ovarian) epithelial tumors.

THE TAKEAWAY

• In the article that accompanies this editorial, Powell et al⁵ present the results of a randomized phase III trial comparing carboplatin/paclitaxel with ifosfamide/paclitaxel in chemotherapy-naive uterine and ovarian carcinosarcoma. The results support that carboplatin/paclitaxel is noninferior and less toxic and should become the preferred regimen for upfront treatment of uterine carcinosarcoma.

In the article that accompanies this editorial, Powell et al⁵ present the results of a randomized phase III trial comparing the standard carboplatin/paclitaxel (CP) regimen typically used to treat epithelial gynecologic tumors with ifosfamide/paclitaxel (IP), a combination developed specifically for sarcomas. Designed as a noninferiority study, protocol GOG-261 published herein not only targeted chemotherapy-naïve patients with UCS but also allowed enrollment of women with ovarian carcinosarcoma, an even rarer tumor type, as a separate group.⁵ The primary study end point was overall survival (OS) for the UCS cohort. Among the 449 evaluable patients with UCS, the CP regimen was found to be not inferior to IP with the median OS of 37 versus 29 months, respectively (hazard ratio [HR], 0.87; 90% CI, 0.70 to 1.075). In addition, CP led to significantly improved progression-free survival (PFS 16 v 12 months; HR, 0.73; P < .01) for patients with UCS. Among the 90 patients with ovarian carcinosarcoma, the differences in survival were similar to those observed in the UCS cohort (OS 30 v 25 months and PFS 15 v 10 months), but not statistically significant because of the much smaller sample size.⁵ Importantly, 30% of women enrolled in this trial were Black, reflecting the disproportionate representation of aggressive UCS among African Americans.⁶

The evolution of treatment strategies building up to the new standard of care established by the GOG-261 trial presented in this issue of the journal starts with the initial observations that ifosfamide was active in UCS, inducing response rates of 35%,⁷ higher than those observed with other single agents tested, including cisplatin (18%),⁸ paclitaxel (18%),⁹ doxorubicin (10%),¹⁰ and etoposide (6.5%).¹¹ Those results laid the foundation for the first phase III randomized trial (GOG 108) that compared the activity of 5-day ifosfamide monotherapy with that of ifosfamide in combination with cisplatin. This trial demonstrated a slight prolongation of PFS (HR, 0.73; P = .02), but no benefit in OS (HR, 0.80; P = .071) for the combination. In addition, the ifosfamide-cisplatin regimen was associated with significant toxicity, including six treatment-related deaths.¹² A subsequent phase III randomized trial (GOG-161) compared the effects of ifosfamide monotherapy with those of IP. Addition of paclitaxel to ifosfamide induced clear benefits in PFS (HR, 0.71, 95% CI, 0.51 to 0.97) and OS (HR, 0.69; 95% CI, 0.49 to 0.97),¹³ rendering IP the preferred regimen for UCS. However, because of the high hematologic and neurotoxicity associated with this regimen, requirement for growth factor usage, and cumbersome multiple day dosing of ifosfamide, it became important to find an equally effective, but less toxic regimen. With this goal, the single-arm, phase II GOG-232B trial evaluated the efficacy and toxicity of CP in women with recurrent UCS. In that study, the response rate to computed tomography was 54% and the median PFS was 8 months,¹⁴ resembling those achieved with ifosfamide-based regimens and leading to its subsequent inclusion for comparison with IP in the current trial.

Found to be noninferior to IP, yet easier to administer in the outpatient setting and harboring a significantly safer toxicity profile, CP has become the new standard for patients with UCS. The IP regimen had required universal use of growth factor support, whereas only two patients randomly assigned to CP received growth factors.⁵ Patients in the IP arm were more likely to experience confusion and grade ≥ 3 neurologic toxicities, known and significant adverse events of ifosfamide.^12,15 All-grade renal toxicity and genitourinary hemorrhage were also more frequent with IP than with CP. By contrast, grade 3 and higher hematologic toxicity was more common in the CP arm, in the absence of growth factor support, but was reversible, anticipated, and not associated with significant clinical complications.⁵ Thus, going forward, CP, the less toxic combination of the two tested, with a comparable level of activity, becomes the preferred regimen for gynecologic carcinosarcoma.

The high level of activity induced by CP suggests that the epithelial carcinoma component of UCS is the main tumor driver, contrary to previous theories speculating that UCS represents either the collision of adjacent carcinoma and sarcoma elements or the divergence of a common progenitor into two elements.¹⁶ Indeed, recent genomic, transcriptomic, and epigenomic studies suggest that UCS shares many common mutations (TP53, PIK3CA, PTEN, FBXW7, PPP2R1A, ARID1A, and CDH4) with uterine carcinomas.^4,17 On the basis of its mutational and transcriptomic signature, UCS resembles either the endometrioid or the serous-type uterine carcinoma, as defined by the TCGA.^3,18 Rare POLE mutations, microsatellite instability, or apolipoprotein B mRNA-editing enzyme, catalytic polypeptide signatures, encountered in uterine cancer, have also been recorded, albeit rarely in UCS.^3,4 A subclassification into the same four genomic subtypes recognized for endometrial cancer: POLE-mutated, MSI-high, copy number low, and copy number high, could be reconstituted for UCS on the basis of its molecular architecture,³ with most serous-like tumors falling into the copy number high category. An epithelial to mesenchymal transition (EMT) transcriptional signature governed by high expression levels of members of the miR-200 family, derepressed epigenetically, and of key transcription factor regulators of EMT (Zeb and Snai) is uniquely characteristic of UCS and may represent a therapeutic target.^3,4 Mutations in genes encoding histones H2A and H2B have been reported in UCS as a distinct feature, and forced expression of the mutant histones was shown to induce sarcomatous differentiation of serous uterine carcinoma cells,¹⁷ further supporting the concept that UCS has an epithelial origin and that epigenetic factors contribute to the sarcomatous transformation. Exome and targeted gene sequencing of adjacent epithelial and sarcomatous elements in the same tumor permitted reconstruction of phylogenetic trees to trace the origin of UCS.³ The clonal analysis suggested that the carcinoma and sarcoma components share driver mutations,³ supporting that the two elements have a common origin. Thus, the newly deciphered mutational and transcriptomic landscape of UCS^3,4,17 coupled with the demonstration in the current trial that a regimen specifically designed to treat epithelial tumors (CP) is highly active in this rare tumor steers for the first time the treatment for UCS toward new directions (Fig 1).

FIG 1.

Model illustrates putative origin of carcinosarcoma in an epithelial progenitor with late dedifferentiation to sarcoma. Common mutations and drivers of EMT are shown. Illustration was created in BioRender. CP, carboplatin/paclitaxel; EMT, epithelial to mesenchymal transition; TGF, transforming growth factor.

First, CP is a backbone on which addition of new targeted agents can be easily built on in the same way that trastuzumab was added for human epidermal growth factor receptor 2 neu–amplified uterine serous carcinomas¹⁹ and bevacizumab for ovarian cancer.²⁰ Perhaps an agent targeting the EMT signature, when developed, an epigenetic modulator or an immune checkpoint inhibitor will find its way into the standard treatment for UCS. Second, the demonstration that CP induces the same benefits in UCS as in endometrial cancer supports a one-size-fits-all strategy and opens the door toward considering UCS under the same umbrella of new clinical trials enrolling patients with high-grade endometrial cancer. The pace of drug development for UCS has been slow until now, and the change in treatment paradigms has lagged when compared with the advances made for other endometrial cancer subtypes. On the basis of the considerations highlighted above, perhaps the time has come for inclusion of patients with UCS into novel therapeutic trials targeting the broader category of uterine cancer. This approach would accelerate the path to approval of new agents for this aggressive disease and fulfill a highly unmet need. In that regard, we welcome the design of the randomized phase III GOG-3031 trial, which will evaluate CP plus dostarlimab versus CP plus placebo (ClinicalTrials.gov identifier: NCT03981796) in patients with recurrent or primary advanced endometrial cancer, including those with UCS. A path to inclusion has been charted.

AUTHOR CONTRIBUTIONS

Conception and design: All authors

Data analysis and interpretation: Daniela Matei

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Paclitaxel and Carboplatin for Uterine Carcinosarcoma: A Path to Inclusion

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).