This study describes the development and validation of a prognostic nomogram to predict recurrence-free survival (RFS) following the resection of localized, primary GIST. A nomogram that assigns predictions for 2-year and 5-year RFS based on tumor size, site of origin, and mitotic index was created based on a series of 127 patients from a single institution. The nomogram was shown to have better predictive accuracy as determined by concordance probabilities than 2 commonly used staging systems developed at the United States National Institutes of Health (NIH) GIST Workshop in 2001. The nomogram had a concordance probability that was higher but not statistically different than that of a third staging system (a 2006 modification of one the NIH staging systems). Nomogram predictions appeared better calibrated to actual RFS that those of this third staging system. The nomogram may be useful to select patients for adjuvant imatinib therapy.
The ability to predict the likelihood of postoperative recurrence for any primary cancer that is treated by surgical resection is important for several reasons. Foremost, patients can be counseled appropriately regarding their likely outcome. If effective adjuvant therapy exists, patients can be selected properly for postoperative treatment. Furthermore, physicians can determine the type (e.g., physical examination, blood tests, or radiologic tests) and frequency of postoperative surveillance for tumor recurrence. The aim of the present study was to establish a prognostic tool to predict RFS for individual patients after complete resection of localized, primary GIST in the absence of adjuvant treatment.
Tumor recurrence is a common event for patients with GIST as RFS ranged from 63 ±4.8% to 78 ±3.5% at 5 years in the 3 datasets in this study (). Mitotic index and size are the most well validated prognostic variables for determining the likelihood of recurrence after complete surgical resection of GIST. We found on multivariate analysis of the MSKCC patients that mitotic index ≥5 was the dominant predictor of RFS (hazard ratio 14.6, p<0.001).27
In contrast, tumor size ≥10 cm had a hazard ratio of only 2.5. Primary tumor site has also been shown to influence outcome in several large retrospective studies.
The 2 NIH Workshop staging systems were developed empirically based on tumor size and mitotic activity with or without primary tumor site ().14,15
Neither had been subjected to statistical validation prior to publication. Subsequently a modification of the NIH-Miettinen staging system has been proposed based on observations in a large number of GIST patients (), but similarly not subjected to statistical validation prior to publication.18
The NIH-Fletcher Staging System has now become the most well studied staging system for GIST.14
The high risk group of that staging system has been reliably associated with an increased risk of recurrence in several reports.20,21,28,29,36,37
It has been noted, however, that the very low risk and low risk groups do not discriminate risk of recurrence.20,21,29
Furthermore, as the high risk group in most studies has a 5-year RFS of approximately 45–50% and often accounts for approximately 50% of the total number of patients,16,20,21,28,29,37
some authors have noted the need for a grouping of very high risk patients.20,21
Our data corroborate these findings (). The very low risk and low risk groups both identify patients with a good prognosis; the majority of patients have a nomogram predicted 5-year RFS of 90–100%. The intermediate and high-risk groups each appear to identify a group of patients with very heterogeneous outcomes, including a large number of patients with nomogram predicted 5-year RFS of 90–100%. Other proposed staging systems, including the one originally proposed by Miettinen15
and its subsequent modifications,18,19
have not been as rigorously evaluated. Likewise, no staging system has been assessed for its ability to assign a quantitative risk of recurrence for individual patients.
In general, prognostic nomograms are better able to predict the likelihood of events for individual patients than staging systems that stratify patients into a few broad groups. Nomograms are based on statistical models that utilize a combination of prognostic variables to determine the likelihood of a certain event. For instance, nomograms for outcome following the resection of gastric24
adenocarcinoma are more accurate in predicting disease-specific survival (DSS) than the corresponding American Joint Committee on Cancer (AJCC) staging systems and these findings have been validated on external data sets.38,39
In this study, we created a nomogram based on patients treated at MSKCC to predict the risk of recurrence after complete resection of localized, primary GIST. The nomogram is shown to predict an accurate relative risk of recurrence as validated by 2 independent series of patients. The concordance probability of the nomogram was very acceptable (0.76 and 0.80 on the validation cohorts). For comparison, the MSKCC pancreas adenocarcinoma and gastric adenocarcinoma DSS nomograms have concordance probabilities of 0.80 and 0.64, respectively, on the original datasets,24,25
and 0.77 and 0.62, respectively, on the validation cohorts.38,39
The nomogram was better able to predict the relative risk of recurrence than the 2 commonly used NIH Workshop GIST stratification schemes. The predictive accuracy of the nomogram as determined by the concordance probability was at least as good as that of the AFIP-Miettinen staging system. The present nomogram can assign numeric predictions for the risk of recurrence at 2 years and 5 years. These predictions appear to be accurate in the 3 cohorts presented. In contrast, while predictions for recurrence could be assigned to the AFIP-Miettinen system stages, these predictions, particularly those of the “high malignant potential” group, were not as well calibrated as those of the nomogram. The difference in the observed RFS of the “high malignant potential” risk group between the 3 datasets may be related to the variable mix of heterogeneous patient outcomes that is seen when nomogram predictions are plotted for this risk group (). Based on the datasets in this study, the AFIP-Miettinen staging system defines 2 stages with very good outcomes, one stage with a good outcome, and one stage with a poor outcome. The 5-year RFS by stage was 100% for the “very low, if any malignant potential” group, 80–100% for the “low malignant potential” group, 69–87% for the “intermediate malignant potential” group, and 8–43% for the “high malignant potential” group. Furthermore, the The AFIP-Miettinen staging system is limited in that it can only assign patients to these broad groups. In contrast the nomogram can calculate risk of recurrence for any individual patient. If used for stratification, the nomogram offers flexibility in defining risk groups. For instance, if it is decided to administer adjuvant treatment to those patients with < 75% 5-year RFS or < 50% 5-year RFS, these groups can be defined.
It is possible that the prognostic value of the nomogram could be improved with the incorporation of additional variables. There have been conflicting results about whether KIT
mutation status affects outcome in resected localized, primary GIST.27,28,36,40–52
The discrepancies may be due in part to differences in how mutation status is analyzed (e.g., presence or absence of KIT
mutation, exon of mutation, or type of mutation). In the multivariate analysis performed on the cohort used to construct the nomogram, mutation status was not an independent predictor of RFS regardless of how it was analyzed.27
Also, we failed to observe an improvement in the accuracy of the nomogram when mutation status was included. The effect of mutational status on the performance of the nomogram in the validation cohorts was not assessed. Ki-67 staining by immunohistochemistry17,53–57
, p16 staining,58
and tumor cellularity28,56
have also been reported to predict independently recurrence in large series of GISTs. One group has proposed a risk stratification system based on Ki-67 staining and tumor size.16
We did not assess the prognostic value of these variables. While the addition of other variables may improve the prognostic ability of the nomogram, the appeal of the current nomogram is that unlike mutational status, Ki-67 staining, and p16 staining, the variables of tumor size, location, and mitotic index are routinely reported by many pathologists and therefore the nomogram should be broadly applicable. Tumor rupture has been described as an adverse prognostic variable.59
Rupture was an infrequent event in the series used to create the nomogram, and thus the association between rupture or spillage and recurrence did not reach statistical significance.
It also should be noted that of the 3 variables used to construct the nomogram, tumor size and mitotic rate might not be measured uniformly across institutions. Tumor size could potentially be influenced by when the specimen is measured in relation to fixation. Mitotic rate is especially subject to variability as it requires the subjective determination of an individual observer about whether an individual cell is undergoing mitosis. Mitotic rate in this study was determined by expert soft tissue pathologists. These variables, nevertheless, appear to be the most important predictors of recurrence in several studies. Despite the potential variability in size and mitotic rate, the nomogram predictions were well calibrated in the 3 datasets in this study.
The standard of care for localized, primary GIST after surgical resection has changed recently based on the results of the ACOSOG Z9001 trial, which showed an increased 1-year RFS in patients assigned to 1 year of imatinib versus placebo.10
The trial was powered on RFS of the entire study population, which was patients with ≥3 cm tumors. Nevertheless, ad hoc
analysis of tumor size (which was the only stratification factor) demonstrated significant differences in RFS between the imatinib and placebo arms in each size category (i.e., 3–6 cm, 6–10 cm, and ≥10 cm). Retrospective analysis of mitotic index determined by central pathologic review and tumor location are underway. Once there is additional follow-up and more events, it will be desirable to validate further the nomogram using the patients assigned to the placebo arm in the ACOSOG Z9001 trial. It is likely that patients at low risk of tumor recurrence do not need adjuvant imatinib. On the other hand, patients at high risk of relapse may require longer periods than 1 year of postoperative therapy.
In summary, we report a nomogram to predict RFS for patients with completely resected localized, primary GIST who have not received adjuvant treatment. This tool can provide an accurate prediction of the risk of recurrence for individual patients as validated in 2 external datasets. The nomogram may prove useful in patient care, interpretation of clinical trial results, and selection of patients for adjuvant therapy.