This article is more than 5 years old. Information may no longer be current.
Nomograms predict risk for tumor progression, death in imatinib-treated patients with GIST
Click Here to Manage Email Alerts
Nomograms effectively predicted the risks for gastrointestinal stromal tumor progression and death and may be valuable for risk stratification and patient counseling, according to study results.
David Goldstein MBBS, FRCP, FRACP, professor in the department of medical oncology at the Prince of Wales Hospital in Sydney, Australia, and colleagues, developed the nomograms as part of a study to stratify patients and to form individual OS and PFS predictions for patients with metastatic gastrointestinal stromal tumors (GIST) who received imatinib (Gleevec, Novartis).
“Our study was a multinational attempt to provide the practicing oncologist who looks after patients with metastatic GIST some guidance of likely outcomes for patients at first presentation with advanced disease,” Goldstein told HemOnc Today.
Researchers evaluated data from a training cohort composed of 330 patients who participated in a multi-institutional, international, phase 3 study that compared a daily dose of 400 mg vs. 800 mg imatinib to develop the nomograms.
Those nomograms were validated using data from a validation cohort of patients undergoing routine treatment for metastatic GIST at one of six major international institutions between 2000 and 2013.
Risk factors for OS and PFS were the size of the largest metastasis, the genotype of the tumor, the mitotic count of the primary tumor and the hemoglobin and blood neutrophil count at the beginning of the imatinib treatment.
Researchers calculated c statistics — a logistical regression measure of how well a model or nomogram can be used to discriminate between subjects having or not having a certain event —of 0.75 for OS in the training group and 0.62 for OS in the validation group. The c statistics for PFS was 0.69 in the training group and 0.62 in the validation group.
In the validation cohort, the nomograms were able to differentiate high- and intermediate-risk patients from low-risk patients. Compared with the low-risk cohort, patients in the high-risk cohort demonstrated a 3.8-fold increase in the risk for death (HR = 3.83; 95% CI, 1.71-8.56), and patients in the intermediate-risk cohort demonstrated a 2.5-fold increase in the risk for death (HR = 2.48, 95% CI, 1.12-5.5).
The risk for disease progression or death also was increased in the high-risk cohort (HR = 2.84; 95% CI, 1.66-4.87) and intermediate-risk cohort (HR = 1.45; 95% CI, 0.87-2.41) compared with low-risk patients.
“Our findings are very practical and can be derived without the need for complex investigations,” Goldstein said. “We found that patient prognosis can be predicted from a combination of the largest metastasis diameter, neutrophil count, primary tumor genotype, hemoglobin and the mitotic count of the primary tumor.”
Goldstein added that the results were first derived from a study in Australia and then further developed using patient data sets from the USA, Poland, Finland and the EORTC. The strength of the analysis lies in the use of data sets from both regular clinical populations as well as a major clinical trial.
“Nomograms offer an alternative to current practice, where estimates of prognosis rely on individual clinician experience or published median survival times,” the researchers wrote. “Another popular alternative would be to use single prognostic factors, such as GIST tumor genotype, or a simple summation of factors to predict good versus poor outcomes… [However], nomograms provide more accurate estimates by combining clinical predictors into single summary measures.” – by Anthony SanFilippo
For more information:
David Goldstein, MBBS, FRCP, FRACP, can be reached at Prince of Wales Hospital, Cancer Care Centre, Barker Street, Randwick NSW, 2031, Australia; email: d.goldstein@unsw.edu.au.
Disclosure: The researchers report no relevant financial disclosures.
Click Here to Manage Email Alerts