Skip to main content
Premium Trial:

Request an Annual Quote

Genomic Profiling Aids Patients With Cancers of Unknown Primary in Trial, Though Benefit is Limited

NEW YORK – The CUPISCO trial has confirmed that genomic profiling and molecularly guided treatment improves outcomes for patients with cancers of unknown primary, or CUP, according to data researchers presented at the European Society for Medical Oncology Congress this week.

However, the benefit seen in the trial with molecularly informed treatment, while statistically significant, was relatively small.

CUP represents between 2 percent and 5 percent of all cancers. Patients present with metastatic malignancies for which a primary site of origin can't be found despite an extensive work-up. "These patients have terrible outcomes," Linda Mileshkin, director of medical oncology at Australia's Peter MacCallum Cancer Center, said, estimating that the median overall survival for patients with CUP is less than a year with standard platinum-based chemotherapy.

In CUPISCO, patients with newly diagnosed CUP were randomized to receive either routine platinum-based chemotherapy or targeted therapy and cancer immunotherapy guided by comprehensive genomic profiling.

In presenting the data from the trial at the meeting, Mileshkin said that CUPISCO demonstrated the value of using comprehensive genomic profiling at the time of diagnosis in CUP patients to allow physicians the opportunity for precision care strategies that could improve outcomes.

Previous studies had shown that actionable abnormalities can be found in about a third of CUP cases. In CUPISCO, investigators hoped to prove that identifying patients with targetable alterations and personalizing their treatment based on that biomarker data would lead to improved survival.

In order to partake in the study, patients had to have all their prior diagnostic work-ups reviewed by an eligibility panel to confirm that past assessments had failed to identify the origins of their cancers. Mileshkin said that this actually led to a striking rate of ineligibility, with more than half of the patients initially recruited to the trial failing this upfront screening.

"One of the reasons for this is that the central eligibility review actually changed the diagnosis of many patients called CUP and said, no, you don't have unfavorable non-squamous CUP, you should go and have normal therapy," said Mileshkin. She and her colleague published a previous report describing this challenge, which she said led ESMO to update its guidelines on the diagnosis and treatment of CUP.

The 631 patients who did pass the stringent assessment process went on to receive three cycles of induction, platinum-based chemotherapy — either carboplatin, paclitaxel, cisplatin, gemcitabine, or carboplatin and gemcitabine. And, while this initial treatment was ongoing, tissue and blood samples, if available, were sequenced by Roche subsidiary Foundation Medicine.

"Over the course of enrollment … from July 2018 to December 2022, the assays actually became more comprehensive and sophisticated and were able to detect more abnormalities, and we incorporated these updates into various protocol amendments throughout the life of the study," Mileshkin said.

At the end of the induction period, study participants were divided into two groups: those who had either a partial or complete response to chemo and those who progressed. The progressing patients could move on to molecularly informed treatment, but the CUPISCO team did not present this data at ESMO.

Mileshkin focused instead on the second group, the chemotherapy responders, who went on to be randomized three to one to receive either molecularly guided therapy based on their genomic sequencing results and the recommendations of a molecular tumor board or to have additional chemotherapy.

Ultimately, patients' oncologists were able to prescribe one of 12 available molecularly informed therapies, some of which were added to the protocol later in the study as new evidence became available. Incyte's Pemazyre (pemigatinib), which last year became available in the US for treating FGFR2-altered cholangiocarcinoma, and Merck's immunotherapy Keytruda (pembrolizumab), which can be given in a tissue-agnostic manner when tumors have a high tumor mutational burden, are two examples of drugs added to the protocol.

About 70 percent of patients had no actionable abnormality but could still receive an immunotherapy-chemotherapy combo.

At a median follow-up of 24 months, Mileshkin said her team deemed CUPISCO a positive study. Molecularly guided treatment reduced the risk of progression or death by 28 percent versus the comparator arm. Patients given the opportunity for precision treatment had a median progression-free survival of 6.1 months versus 4.4 months for those who got chemo.

Mileshkin highlighted that the progression-free survival benefit was even better for patients in the molecularly guided therapy arm who actually had an abnormality to target. In this group, the risk of progression or death was 33 percent lower with a median progression-free survival of 8.1 months.

The best confirmed overall response rate, which includes responses after the initial chemotherapy induction period, was higher in the molecularly targeted treatment arm at 18 percent versus 8 percent with chemotherapy. The duration of response was similar in both arms, Mileshkin reported, and there was no real difference in disease control rate, which was about 60 percent for both arms.

Mileshkin said that while overall survival data are still immature in CUPISCO, investigators are already seeing a trend suggesting benefits with molecular testing. At the last data cutoff, median overall survival was 14.7 months in the molecularly guided therapy arm compared to 11 months with chemotherapy. The team plans to reanalyze overall survival in about a year.

Adverse event rates were generally similar or lower with molecularly guided therapy. Patients who experienced adverse events needed dose modifications or had to have their dose interrupted, but the rate of serious events were lower in the molecularly guided therapy arm. "Importantly, at the time of this analysis, there were no study-related deaths in either arm," Mileshkin said.

"Precision oncology can only be realized if we can get access to broad molecular profiling for our patients, as well as the relevant associated treatments," Mileshkin said.

Discussing the results, Federica Di Nicolantonio, an associate professor at the University of Turin School of Medicine who was not involved in CUPISCO, said that an improvement in median progression-free survival of 1.7 months, while statistically significant, is admittedly "not transformative."

Still, she recognized the trial as an achievement, given the failure of previous CUP studies that attempted to direct treatment based on gene expression assays designed to identify tumor origin or tissue type, such as the GEFCAPI trial.

Gene expression testing may yet be vindicated though, Di Nicolantonio said, given that these failed trials were performed prior to the availability of several new targeted therapies and before the broadening of immunotherapy indications. Indeed, another study presented at ESMO showed positive results comparing empiric chemotherapy against treatment informed by a gene expression test developed by Chinese diagnostics firm Canhelp.

In that much smaller trial, researchers randomized 182 patients to either site-specific cancer therapy guided by the Canhelp-Origin test or standard chemo. In the tested arm, nearly three quarters of patients were matched to a site-specific drug while 26.4 percent received a standard chemotherapy regimen. At a median follow-up of 43 months, the site-specific therapy group had a median progression-free survival of 9.6 months versus 6.6 months in the control group.

Some thought leaders in the field have coalesced around the need for both molecular profiling and tumor origin testing for identifying the best treatments for CUP patients, given the fact that many molecularly targeted drugs only benefit patients with specific types of cancer.

Di Nicolantonio also highlighted emerging evidence that artificial intelligence might aid in CUP assessment, citing a recent study in Nature Medicine.

Mileshkin said that she and her colleagues hope that CUPISCO will be a first step for companies like Foundation Medicine and others in building reimbursement for their tests outside the US. Ensuring patients then have access to the therapies that actually improve their survival is an even greater challenge and a "bigger piece of work," she added, but oncologists are hopeful that a growing collection of positive clinical impact data can make a difference.