Genomic Profiling in Sarcoma Shows Potential in Guiding Clinical Management

A new single-center report has found that the results of next-generation sequencing (NGS)–based molecular profiling for non–GI stromal tumor (non-GIST) sarcomas provided information used to effectively guide clinical management.

A team of authors led by Spandana Boddu, a research assistant at Moffitt Cancer Center in Tampa, Florida, analyzed data from 114 patients with a diagnosis of non-GIST sarcoma who underwent molecular profiling during treatment at the center between May 2013 and March 2017.

“We found that clinical management was affected by NGS results in a small but notable percentage of patients, including those with a diagnosis change (4.4%) and/or in whom therapeutic selection was altered by the treating physician because of findings
(13.2%),” the authors wrote in JCO Precision Oncology. “The genes most commonly affected by pathogenic mutations in our cohort mirror those most widely reported in sarcomas to date.”

Boddu et al used Moffitt’s Personalized Medicine Clinical Service (PMCS) database to identify all non-GIST sarcoma patients who underwent commercial genomic testing between May 2013 and March 2017. A PMCS review of all patients with solid or hematologic malignancies who undergo commercial genomic sequencing is standard at Moffitt.

Each patient’s genomic findings were also included in the Clinical Genomic Action Committee (Moffitt Molecular Tumor Board) database. The database included demographic, clinical, and histologic information, such as the type of genomic test and the source of biopsy specimen. Each gene, mutation, and allele frequency or copy number along with microsatellite status and mutation burden, if available, was also recorded in the molecular tumor board database.

Each patient underwent molecular profiling on the same comprehensive genomic panel on a commercial platform. The authors performed a chart review to retrospectively collect treatments and outcomes data and confirm pathology findings.

The authors identified “clinically actionable” genetic alterations on the basis of the assessment of the commercial testing company and the PMCS review. They also defined genetic alterations as clinically actionable if they had been previously documented as “providing diagnostic or prognostic information or to predict response or resistance to commercially available or investigational agents.”

Boddu et al defined actionability as “predicting response to approved drugs available for the patient’s diagnosis (on-label), for another diagnosis (off-label), or investigational drugs being studied in humans for whom the genetic alteration has been shown to serve as a suggested biomarker for response,” they wrote. “Assessment of clinical actionability was limited to mutations previously reported or likely to be driver mutations based on available literature and functional classification and did not consider analysis of variants of unknown significance.”

Of the 114 patients, slightly over half were female (n = 63, 55%). Patients’ median age at diagnosis was 55 (interquartile range [IQR], 38 to 65) years. Most patients had metastatic disease (85.1%) and had received prior systemic therapy (81.6%). A large majority of patients had soft tissue tumors (n = 94, 82.5%), while the remaining 20 patients had bone sarcomas (17.5%).

Patients’ histologies were consistent with the most common sarcoma subtypes, including leiomyosarcoma (16.7%), well-differentiated/dedifferentiated liposarcoma (12.2%), and undifferentiated pleomorphic sarcoma (10.5%). Chondrosarcoma (7.0%) and osteosarcoma (6.1%) were the most common types of bone tumors.

NGS detected 438 genetic variants among the collection of 114 tumors that were presumed to be oncogenic. Almost all tumors had at least one driver variant (96.7%), while the median number of driver variants per tumor was 3 (range, 0 to 19). Regardless of sarcoma subtype, the most common alterations were in TP53 (36.8%), CDKN2A/B (20.2%), CDK4/MDM2 (19.3%), ATRX (13.2%), and RB1 (13.2%). About 60% of alterations were structural, including 157 amplifications and 66 copy-number losses.

No patients displayed evidence of microsatellite instability. NGS was able to assess tumor mutational burden in 106 patients. The clear majority of these patients had <6 mutations (84.9%). Around 13% of patients were characterized as having intermediate tumor mutational burden (6 to 20 mutations), while less than 2% of patients had >20 mutations.

In 5 cases (4.4%) treating physicians viewed the genomic findings as either diagnosis-changing or diagnosis-modifying. Boddu et al described a “low grade spindle cell sarcoma” being reclassified as a desmoid tumor after the detection of a CTNNB1 mutation prompted additional pathologic review.

Additionally, the diagnosis of “poorly differentiated sarcoma, favor neurogenic tumor/MPNST [malignant peripheral nerve sheath tumor]” was revised to synovial sarcoma after NGS detected a typical SS18-SSX2 fusion. This mutation was then confirmed by conventional testing. “In the remaining three cases, novel or seminovel fusions that were felt to be disease-defining were detected in cases of ‘small round cell sarcoma, not otherwise specified’, including one each of EWSR1-PATZ1, BCOR-ZC3H7B, and PHF1-TFE3,” wrote Boddu et al.

NGS testing reports included a therapeutic treatment recommendation for 88 patients (77.2%). Following Moffitt review of the genomic findings, 56 patients (49.1%) were classified as having an actionable result. Most of these patients had options to pursue both off-label targeted therapy and molecularly matched clinical trial options. Nine of these patients found that their only significant alteration was already known or assumed. These findings included MDM2/CDK4 amplification in dedifferentiated liposarcoma or NF1 in malignant peripheral nerve sheath tumor.

NGS findings guided treatment management in 15 cases (13.2%). This includes patients for whom a standard treatment was preferred over an alternative treatment and patients for whom chosen therapies changed due to a change in diagnosis based on NGS findings. Boddu et al reported that 4 of 15 (26.7%) NGS-influenced treatment plans resulted in clinical benefit such as partial response or stable disease > 6 months.

Boddu et al concluded that clinical genomic profiling altered the disease course of “a sizeable minority” of sarcoma patients at Moffitt. “Looking forward, we expect the number of patients who have actionable NGS findings to steadily increase as more molecularly targeted therapies become available and as genomics becomes increasingly used as a biomarker for immunotherapy response,” they wrote. “At present, our experience is that patients with sarcoma who are most likely to benefit or at least have their clinical course altered by NGS are those with an unclear diagnosis or rarer sarcoma subtypes in which potential genomic targets are not well known, and patients in whom early-phase trials are being considered.”