GENETICS OF CUP

More than two thirds of patients with cancer of unknown primary (CUP) have mutations detectable on circulating tumor DNA (ct-DNA) testing, or so-called liquid biopsy. For these patients, ct-DNA testing could alter their treatment and improve their outcomes, say UK researchers.

The findings come from a proof of principle study presented here at the European Society for Medical Oncology (ESMO) 2018 Congress.

The study included 25 patients with CUP. Main disease sites varied among the patients. Several patients had previously undergone testing with the Guardant360(G360; Guardant Health) ct-DNA test, which assessed genetic variations across a panel of genes.

The results showed that 68% of tumors were carrying mutations that were potentially treatable with targeted therapies. For two patients, treatments were changed on the basis of the findings.

Kai Keen Shiu, MD, PhD, from the Sarah Cannon Research Institute and clinical lead for the Cancer of Unknown Primary Service at University College London, United Kingdom, told Medscape Medical News that the first thing that clinicians need to do when a CUP patient presents is to determine whether the patient indeed has a CUP.

"Common things are common, so even when you think you've got a true CUP, it's more likely to be a common cancer, like an unusual breast cancer, lung cancer, GI cancer...you just can't find the primary," he commented.

"If you think about it, most of our adult cancers have genetic alterations for which we already use genetic profiling to use targeted therapy. So why would a CUP population be any different?" he asked.

"We know that usually, these undifferentiated CUPs are quite mutated and have a high mutation burden, and therefore, if you've got lots of mutations...then you should be able to target that, either with targeted therapy or now we think with immunotherapy," he said.

CUP patients often have a poor performance status and so cannot tolerate lots of tissue biopsies. This is where a liquid biopsy comes into its own.

"You can do the liquid biopsy to see what the whole genome landscape is," Shiu noted. He added that it is "also not particularly painful intervention for the patient and gives results that you work on."

Shiu said that, having shown the potential of ct-DNA testing in this patient population, he and his colleagues are now planning a phase 2 trial of targeted therapy based on genetic profiling of the tumor for patients with a CUP.

Finding Alterations That Are Potentially Actionable

Approached for comment, Rodrigo Dienstmann, MD, Vall d'Hebron Institute of Oncology, Barcelona, Spain, who was not involved the study, said that "this is not the first study to look at the potential and value of doing next-generation sequence in cancer of unknown primary."

What is "interesting" about this study, he told Medscape Medical News, is that often, there is not enough tissue to be able to perform next-generation sequencing from a tissue biopsy specimen, and the tissue that is available is often of poor quality.

Dienstmann said: "They did find alterations that are potentially actionable in a good proportion of the patients...but with different levels actionability." He noted, "Actionability is always a matter of how you rank them."

He said that it makes more sense to perform genetic testing if the primary tumor is lung cancer, "probably less so if you have what you suspect is a head and neck primary or pancreatic primary."

In these situations, he believes that there is probably greater value in assessing the pathology of the tumor, the immunohistochemistry, and the clinical symptoms.

On the other hand, he noted, a clinical presentation such as that of a smoker with adenocarcinoma who has lots of lung lesions as well as cancer of the liver and lymph nodes "is telling me something...and obviously there is a high value of next-generation sequencing."

Dienstmann summarized: "Overall, it's good results, and this is a subgroup of patients that probably deserves next-generation sequencing, in my opinion.

"Whether this technology is better than tumor testing and looking at larger gene panels, tumor mutation burden, microsatellite instability, and other aspects, we don't know, but this study shows that it is feasible, and apparently it does identify alterations that you can act upon."

Details of the Study

The researchers note that up to 9% of all cancer cases present as metastatic CUP. Despite multiple guidelines having recommended multiagent cytotoxic chemotherapy, outcomes are poor, with a 5-year overall survival of just 11%.

Previous studies have indicated that 96% of CUP cases have at least one genetic alteration and that 85% of tumors have potentially clinically relevant genomic alterations. These include RTK/RAS signalling pathway alterations in 72% of adenocarcinoma cases and in 39% of nonadenocarcinoma CUP cases.

To determine whether ct-DNA testing could identify actionable gene alterations in CUP, the team studied patients referred to the Sarah Cannon Research Institute for analysis with the G360 test.

This panel, which covers all National Comprehensive Cancer Network somatic mutations, examined single-nucleotide variants in 73 genes, gene copy number amplifications in 18 genes, fusions/rearrangements in six genes, and indels in 23 genes.

To qualify for inclusion, the patients had to truly have CUP, and their cases had to have been discussed at a CUP multidisciplinary team meeting.

In all, 25 patients (median age, 67 years) were recruited from 2017 to 2018. Just more than half (14) of the patients were women.

The main disease sites were lymph nodes (n = 8), the pelvis and peritoneum (n = 8), liver (n = 6), bone (n = 3), adrenal gland (n = 2), and brain (n = 1).

The majority of patients were treatment naive at study entry. Eight had received one line of therapy, three had received two lines, one patient had undergone three, and another four.

The median turnaround time for ct-DNA testing was 10 days (range, 6 - 15 days).

Four patients had no detectable mutations, which the researchers say may have been related to their treatment.

However, 17 (68%) patients had potentially actionable mutations. These consisted of two BRAF-V600E mutations and five KRAS mutations, as well as FGFR, KIT, PIK3CA and ERRB2 mutations. In one patient, MYC amplifications were present.

Twelve patients had three or more somatic mutations, including variants of unknown significance. Six patients had six or more mutations.

Two patients with detectable mutations went on to receive targeted therapy within 8 days of testing.

In one case, a 62-year-old woman who had been diagnosed with intrahepatic cholangiocarcinoma and who had had a mixed response and poor tolerability to chemotherapy was found to have a BRAF-V600E mutation.

BRAF/MEK inhibitor therapy resulted in clinical improvement. There was a significant reduction in right-sided pleural effusion, a reduction in the size and number of metastases, and a reduction in her liver mass.

The team concluded that "the presence of actionable targets supports the inclusion of patients with CUP into molecularly driven clinical trials."

The study was supported by Guardant Health, which provided the 360TM test. One author is an employee of Guardant Health. The other authors have disclosed no relevant financial relationships.

European Society for Medical Oncology (ESMO) 2018 Congress. Abstract 152P, presented October 20, 2018.