Sarcomas are a heterogeneous and motley collection of cancers, which struggle with an identity crisis on many levels. The trials often lump vastly different subgroups, and are often unable to collect sufficient numbers of any one disease subtype to complete a unique cohort.
The first set of posters looked at genomic analyses of sarcomas by The Cancer Genome Atlas (TCGA). The TCGA started off with different subgroups, but ultimately was condensed into one sarcoma working group in order to facilitate forward progress of this important project. Thus, the TCGA is profiling a smattering of many types of sarcomas.
The first poster discussed was the TCGA preliminary whole exome sequencing results. P53 was frequently mutated (which is not a surprise within sarcomas). ATRX was another commonly altered gene. ATRX is involved in chromatin remodeling and telomere maintenance. In cancer, it functions as a tumor surpressor and loss is associated with genomic instability. Mutations in ATRX are associated with tumors that use the ALT pathway.
Next generation sequencing of 26 synovial sarcomas was the focus of the next poster. This was accomplished via a hotspot panel using an Ion Torrent Personal Genome Machine. Consistent with what is published on the disease, these synovial sarcomas painted a pretty quiet genomic picture. The two notable genes that were mutated were CCND1 and KRAS. Interestingly, the case that had a CCND1 mutation also stained strongly positive by immunohistochemical (IHC) staining, suggesting it may actually be a driver mutation in that particular tumor.
The authors also looked at a subset of cases for copy number variation (CNV), and described differences in gain and losses between primary and metastatic lesions. Perhaps most interesting was the discordant loss of chromosome arm 6q which was observed in one, but not another metastatic lesion from the same patient, and correlated with response/progression (respectively) of disease in a patient on pazopanib (Votrient). Therefore, 6p loss may be a potential biomarker for pazopanib and warrants further investigation; however, there was concordance of mutations between primary and metastatic disease. Finally, they concluded that in a relatively quiet genomic landscape that this further lends support that canonical SYT-SSX gene fusion is sufficient for tumor formation.
Prognostic gene expression in rhabdomyosarcoma (RMS) was the next poster highlighted in the tour de sarcoma genome. RMS can be classified by molecular analysis as PAX-FOXO gene fusion-positive (worse prognosis) and negative (mixed prognostic group). The authors proposed a five-gene signature for prognosis (MG5 score). The Kaplan-Meier curve prognosis was much worse for the MG5 high group than the MG5 low group. Could this be used in the same manner that prognostic gene signatures are used in other diseases such as breast cancer?
The next series of posters moved on to clinical trials. The first poster reported the overall survival (OS) in EURAMOS-1 osteosarcoma patients, which is an international, randomized controlled trial to improve outcomes in osteosarcoma therapy. The 5-year OS was 71% (95%CI: 69%-73%).
The second poster reviewed the SARC011 trial, a phase II trial of R1507, a recombinant human monoclonal antibody to the insulin-like growth factor-1 receptor (IGF-1R inhibitor) for the treatment of patients with recurrent or refractory Ewing sarcoma, osteosarcoma, synovial sarcoma, and RMS. Early PET scan (day 9) seemed to be predictive of response, but the discussant raised the important question if this is applicable to just this agent (an IGF-1R inhibitor) or other therapies as well.
The next poster looked at 40 years of randomized controlled trials in soft tissue sarcomas (STS). Findings included the fact that only 54% of studies defined the primary endpoint clearly; 14% suggested benefit while primary endpoint was not met. In 47%, toxicity was comprehensively reported while 14% did a poor job reporting (this did not change over time). There is a reasonable correlation between PFS and response rate (RR) as surrogates for OS. However, the discussant reminded us that we should be cautious in attempting to use RR as a surrogate for OS for noncytotoxic therapy since RR to targeted therapies may be low, but still translate into an improvement in OS.
Finally, we heard about early-phase clinical trials for STS. The first study was a phase Ib dose escalation of TRC105, which is a chimeric IgG1 monoclonal antibody that binds CD105 (endoglin). Endoglin is a cell membrane glycoprotein receptor for TGFβ and bone morphogenetic proteins on endothelial cells. It is an antiangiogenic target implicated in VEGFR resistance. Therefore, it may have direct tumor effect, stromal effect, and overcome VEGFR resistance. It was well-tolerated with primarily low-grade toxicities and most notably included epistaxis and gum bleeding.
The next poster looked at a phase II study of tivozanib (AV-951) in patients with pretreated STS. Fifteen out of 58 patients had prior pazopanib, some of which had activity. Subjectively, patients seemed to tolerate tivozanib better than pazopanib.
Finally, we heard about belinostat (Beleodaq), which is a HDAC inhibitor that is approved for peripheral T-cell lymphoma. Here the study was of a phase Ib combining belinostat with doxorubicin. The dose escalation included other solid tumors, but the maximum tolerated dose expansion was specific to STS. The outcome to determine optional expansion was not met since responses were seen after the first two cycles. The presence of four chondrosarcoma patients in a small sample size may have limited its chances since chondrosarcomas are notoriously unresponsive to anything. That said, combined with a prior negative study of doxorubicin with a different HDAC inhibitor, this result questions as to whether there is a future with HDAC inhibitors with doxorubicin. The reviewer left us with a nice case study of a complete response of a cutaneous angiosarcoma to TRC105.
During the discussion, a question was posed about the amount of input from the investigator versus the industry partner. Most investigators from the presented posters confirmed that the investigators had substantial discussion, but primarily found the collaboration very collegial and beneficial. As a side note, the industry intent to use traditional standard of care options such as doxorubicin as a backbone, I suspect, is really based on the FDA requirements if it were to go forward into a registration-type setting.
The final group of posters looked at cfDNA testing in patients with gastrointestinal stromal tumors (GIST), a rare type of sarcoma found in the wall of the stomach. This holds the potential to be very powerful since it provides the opportunity for a liquid biopsy, which is a noninvasive blood test. This allows evaluation of serial biopsies and also observation of evolution of the mutational profile over time.
In summary, there has been modest progress in the area of sarcoma, but much remains to be answered in such a heterogeneous collection of diseases.