Low signature value ID’d those to benefit most from adjuvant chemotherapy

In patients with gastric cancer, a newly constructed, tumor-associated collagen signature (TACSGC) of the tumor microenvironment was significantly associated with prognosis, a retrospective cohort study found. In addition, a low TACSGC predicted survival benefits from adjuvant chemotherapy in patients with stage II and III gastric cancer.

In a training cohort made up of 294 consecutive patients, disease-free survival (DFS) at five years was 71.7% (95% CI, 65.7%-78.2%) in patients with a low TACSGC value versus only 25.9% (95% CI, 18.4%-36.3%) in patients with a high TACSGC, Dexin Chen, MD, Southern Medical University, Guangzhou, People’s Republic of China, and colleagues reported in JAMA Network Open.

Overall survival (OS) at five years was 74.7% (95% CI, 68.9%-81%) in patients with a low TACSGC level compared to 29.2% (95% CI, 21.4%-39.8%) in those with a high TACSGC, investigators added. Results were further verified in the validation cohort which consisted of 225 patients.

Finally, patients with stage II and III gastric cancer with low TACSGC levels also had a more favorable response to adjuvant chemotherapy than those with high TACSGC levels, the authors noted.

“Adjuvant chemotherapy is the standard treatment for patients with stage II and III GC to improve survival outcomes. However, many patients with stage II and III GC do not experience survival benefits from adjuvant chemotherapy, indicating that some patients might be overtreated because their tumors are not sensitive to the given type of chemotherapy,” wrote Chen and fellow researchers.

“[T]he [current] pathological staging system provides inadequate prognostic information and fails to accurately identify which patients might benefit from adjuvant chemotherapy,” the researchers wrote and concluded, “[These] findings suggest that TACSGC provides additional prognostic information for patients with GC (gastric cancer) and may distinguish patients with stage II and III disease who are more likely to derive benefit from adjuvant chemotherapy.”

In total, the study included 519 patients at a median age of 57 years (interquartile range [IQR], 49-65 years), almost 70% of whom were male. The training cohort included 294 patients from the Nanfang Hospital, Southern Medical University, People’s Republic of China, and the validation cohort included 225 patients from the Fujian Provincial Cancer Hospital, Fujian Medical University, also in the People’s Republic of China. All patients had undergone radical gastrectomy and had at least 15 lymph nodes harvested at the time of surgery.

Using multiphoton imaging, researchers extracted 146 collagen features in the tumor microenvironment. As the main component of the extracellular matrix, collagen accounts for most of the matrix’s functions. The extracellular matrix, in turn, is important for regulating neoplastic progression and response to chemotherapy.

“Multiphoton imaging has been recognized as a novel method for optical biopsy of samples.17 Because of the comparable results between multiphoton imaging and hematoxylin and eosin staining, pathologists or clinicians could differentiate the cancerous and normal tissues of the stomach after minimal training,” the researchers wrote.

A TACSGC that included nine collagen features was subsequently constructed from multiphoton imaging, and an optimal cut-off value of 2.81 was set to separate patients into high versus low TACSGC groups.

“The TACSGC level was significantly associated with tumor differentiation, depth of invasion, lymph node metastasis, and distant metastasis,” Chen and colleagues noted. “[And p]atients with higher TACSGC values were more likely to experience disease recurrence or death.”

Importantly, no significant differences were found in the TACSGC values between the training and validation cohorts.

Integrated nomograms that included five independent prognostic predictors of DFS and OS, namely carbohydrate antigen 19-9; depth of invasion; lymph node and distant metastasis, and TACSGC were subsequently established for both disease-free and OS.

“Compared with clinicopathological models…the integrated nomograms yielded an improved discrimination for prognosis prediction in a C index comparison,” Chen and colleagues observed.

For example, in the training cohort, the clinicopathological models had a decreased C index of 0.78 (95% CI, 0.67-0.84; P=0.002) for DFS and of 0.80 (95% CI, 0.73-0.86; P =0.03) for OS compared to the integrated nomograms. The C statistic is the ability of the model to distinguish between patients who will go on to have the event of interest and those who will not. Similar reductions in the C indexes for the clinicopathological models were also shown in the validation cohort.

DFS rates were also 35% better for patients with stage II and III gastric cancer who had low levels of TACSGC (HR: 0.65; 95% CI, 0.43-0.96; P=0.03) compared to patients with high TACSGC levels, and OS was 45% longer (HR: 0.55; 95% CI, 0.36-0.82; P=0.004).

“An accurate prediction of prognosis and survival benefits from adjuvant chemotherapy in patients with [gastric cancer] GC is integral to treatment decision-making,” the authors pointed out.

“[B]y incorporating TACSGC into clinicopathological models, we found that the TACSGC could provide additional prognostic information [and]… was a potential indicator of survival benefits associated with chemotherapy in patients with stage II and III GC,” they highlighted.

Commenting on the findings, Jesse Berlin, ScD and Daniel Catenacci, MD, statistical editor and associate editor, respectively, at JAMA Network Open in Chicago, pointed out that the current study showed the protective association between chemotherapy and outcomes was stronger in those with a low versus a high TACSGC, but the extent of the difference between these two groups depended on the stage of gastric cancer.

“For patients with stage 2 disease, there were protective associations, numerically, in the high TACS group, but they were not statistically significant and smaller in magnitude than the associations in the low TACS group,” the editorialists wrote.

For patients with stage 3 disease, the HRs for chemotherapy were close to the null value of 1.0 in the high TACS group, but in the low TACS group, a favorable association in the low TACS group persisted, they added.

Nevertheless, Berlin and Catenacci agree that the ability to predict response to chemotherapy is of value.

“If clinicians can focus chemotherapy on the patients most likely to benefit, they can avoid the toxic effects of chemotherapy in those least likely to benefit,” they observed, but added that whether or not the TACSGC provides unique value in this prediction does require further analyses.

“In summary, Chen and colleagues have done an excellent job of showing that TACS may be a useful addition in helping understand gastric cancer prognosis,” the editorialists noted.

The additional value of this score seems to be relatively small in this particular context and further validation is needed from independent data sets to better judge its value, concluded Berlin and Catenacci.

Study limitations include its retrospective nature, specimen acquisition from only two medical centers in China, and the as-yet-unknown mechanisms of TACSGC in predicting prognosis and benefits from adjuvant chemotherapy.

  1. A newly constructed tumor-associated collagen signature in patients with gastric cancer better predicted disease-free and overall survival compared with the current staging system.

  2. A low value of the newly constructed tumor-associated collagen signature in patients with stage II and III gastric cancer identified which patients would most benefit from adjuvant chemotherapy.

Pam Harrison, Contributing Writer, BreakingMED™

The study was funded by the National Natural Science Foundation of China, among many others.

The authors had no conflicts of interest to declare.

Berlin reported being a former employee of Johnson & Johnson.

Catenacci reported receiving personal fees from Guardant Health, Tempus, Foundation Medicine and Natera.

Cat ID: 120

Topic ID: 78,120,730,188,120,192,925