Study results unveiled Sunday are the clearest sign yet a closely watched cancer medicine from Amgen can reliably shrink tumors fueled by mutations in a gene called KRAS.

For decades, KRAS mutations have been considered unassailable by treatment, resisting drug development advances that have exposed the genetic roots of other cancers. Amgen’s drug, called sotorasib, could become the first therapy to target KRAS, which is often altered in lung, colon and pancreatic cancers.

But the new data, published in The New England Journal of Medicine and presented at the European Society for Medical Oncology’s virtual meeting, also leave room for improvement, suggesting Amgen’s breakthrough may only be a first step in treating KRAS cancers.

Nineteen of the 59 patients with advanced lung cancer who were treated in the trial responded to sotorasib. The rate of response, at 32%, is about two times higher than what would be expected for chemotherapy, but falls short of preliminary numbers previously disclosed by Amgen from fewer participants in the study.

Those early findings caught the attention of cancer doctors, turning sotorasib into the star of Amgen’s drug pipeline and one of the most keenly tracked in oncology. With results from more patients, Sunday’s presentation give a more definitive sense of what can be expected from treatment.

Study participants have now also been followed for longer, allowing Amgen to calculate the time until disease worsening or death.

At a median of 6.3 months, the progression-free survival shown Sunday is greater than the 2 to 4 months that would typically be expected for chemotherapy in previously treated patients. Participants in Amgen’s study had all received chemotherapy in the past, and most had also been treated with immunotherapy, too.

“For later lines of therapy, that’s a very good endpoint,” said Marwan Fakih, an oncologist at City of Hope and an investigator in the trial. The data, he added in an interview, “position this drug post chemotherapy as a very viable option of treatment for these patients.”

Among the 34 lung cancer patients given the highest dose of sotorasib, which Amgen is using in larger studies now ongoing, responses to treatment lasted a median of just under 11 months — well above what some analysts had been expecting.

The rate of response in these patients, 35%, was also slightly higher than in the broader group, which included those given the three lower doses tested.

Gilberto Lopes, an oncologist at the University of Miami’s Sylvester Comprehensive Cancer Center, said the data show sotorasib to be a “welcome addition” to doctors’ arsenal of treatments for lung cancer.

But, he added in an email to BioPharma Dive, “the numbers are indeed a little disappointing when we compare them with targeted therapies to other molecular alterations that have recently been approved,” such as Eli Lilly’s Retevmo.

The progression-free survival for sotorasib, while “very encouraging” for the first drug to target KRAS, “is less than what we would hope,” Lopes said. Lopes, who was not involved in the trial, reviewed the data at BioPharma Dive’s request.

Amgen’s study also included 42 patients with colorectal cancer and 28 with other solid tumors known to harbor KRAS mutations. Treatment didn’t lead to as promising results as in lung cancer, however. Only three colorectal cancer patients responded, as did one each with melanoma, pancreatic, appendiceal and endometrial cancers.

The differences in response, Fakih and his fellow study investigators wrote in NEJM, suggest other cancer-linked genes are mutated more frequently in colorectal cancer. Turning off KRAS, in other words, may be necessary but not sufficient to shrink tumors in these patients.

The results also suggest the same might be true in some lung cancer patients, too. All responses to treatment were considered partial, meaning sotorasib led to reductions in tumor size yet did not eliminate the cancer. And a good number whose disease was considered stable saw their tumors shrink by a small amount, but not enough to be classified as a response.

“That the partial responses [are] a great bit more durable than stable disease tells me there is tumor heterogeneity,” said City of Hope’s Fakih. “This heterogeneity is leading to resistance.”

Some patients, for instance, initially responded only for disease progression to occur soon after.

Combining sotorasib with other drugs could help address this, and potentially improve on the number of patients who respond. Amgen is currently studying the drug together with immmunotherapy and several different types of targeted treatments.

Importantly for those combination efforts, sotorasib is relatively well tolerated as cancer drugs go. Nearly six in 10 patients tested experienced a treatment-related side effect, with 12% considered more severe. Both rates are lower than chemotherapy-containing regimens, Lopes said, and no dose-limiting toxicities were observed.