Sub-regional CT Radiomics for the Prediction of Ki-67 Proliferation Index in Gastrointestinal Stromal Tumors: A Multi-center Study

RATIONALE AND OBJECTIVES

The objective was to assess and examine radiomics models derived from contrast-enhanced CT for their predictive capacity using the sub-regional radiomics regarding the Ki-67 proliferation index (PI) in patients with pathologically confirmed gastrointestinal stromal tumors (GIST).

METHODS

In this retrospective study, a total of 412 GIST patients across three institutions (223 from center 1, 106 from center 2, and 83 from center 3) was enrolled. Radiomic features were derived from various sub-regions of the tumor region of interest employing the K-means approach. The Least Absolute Shrinkage and Selection Operator (LASSO) regression was employed to identify features correlated with Ki-67 PI level in GIST patients. A support vector machine (SVM) model was then constructed to predict the high level of Ki-67 (Ki-67 index >8%), drawing on the radiomics features from each sub-region within the training cohort.

RESULTS

After features selection process, 6, 9, 9, 7 features were obtained to construct SVM models based on sub-region 1, 2, 3 and the entire tumor, respectively. Among different models, the model developed by the sub-region 1 achieved an area under the receiver operating characteristic curve (AUC) of 0.880 (95% confidence interval [CI]: 0.830 to 0.919), 0.852 (95% CI: 0.770-0.914), 0.799 (95% CI: 0.697-0.879) in the training, external test set 1, and 2, respectively.

CONCLUSION

The results of the present study suggested that SVM model based on the sub-regional radiomics features had the potential of predicting Ki-67 PI level in patients with GIST.

Prognostic evaluation of the Ki-67 labeling system in histological grading of non-small round cell sarcoma: a supplementary analysis of a randomized controlled trial, JCOG1306

BACKGROUND

Soft tissue sarcoma (STS) has various histological types and is rare, making it difficult to evaluate the malignancy of each histological type. Thus, comprehensive histological grading is most important in the pathological examination of STS. The Fédération Nationale des Centres de Lutte Contre le Cancer (FNCLCC) grading system is most commonly used in daily pathological analysis of STS. Among the FNCLCC grading system parameters, mitotic count is a key morphological parameter reflecting the proliferative activity of tumor cells, although its reproducibility may be lacking. Here, we compared the prognostic utility of the conventional and modified FNCLCC grading systems in JCOG1306.

METHODS

We analyzed 140 patients with non-small round cell sarcoma. We performed Ki-67 immunostaining using open biopsy specimens before preoperative chemotherapy in all patients. We assessed histological grade in individual cases by conventional FNCLCC grading (tumor differentiation, mitotic count, and necrosis) and modified FNCLCC grading using the Ki-67 labeling index instead of mitotic count. We conducted univariable and multivariable Cox regression analyses to investigate the influence of grade on overall survival.

RESULTS

In univariable analysis, prognosis was worse for patients with conventional FNCLCC Grade 3 tumors compared with Grade 1 or 2 tumors (hazard ratio [HR] 4.21, 95% confidence interval [CI] 1.47-12.05, P = 0.008). Moreover, prognosis was worse in patients with modified FNCLCC Grade 3 tumors compared with Grade 1 or 2 tumors (HR 4.90, 95% CI 1.64-14.65, P = 0.004). In multivariable analysis including both conventional and modified FNCLCC grading, the modified grading more strongly affected overall survival (HR 6.70, 95% CI 1.58-28.40, P = 0.010).

CONCLUSIONS

The modified FNCLCC grading system was superior to the conventional system in predicting the prognosis of patients with non-small round cell sarcoma according to this supplementary analysis of data from the randomized controlled trial JCOG1306.

Prognostic usefulness of a modified risk model for solitary fibrous tumor that includes the Ki-67 labeling index

Background

Predicting the prognosis of patients with solitary fibrous tumor (SFT) is often difficult. The prognostic risk models developed by Demicco et al. are now the standard for evaluating the risk of SFT metastasis in the current World Health Organization classification of soft tissue and bone tumors.

Methods

In this study, we examined the prognostic usefulness of a modified version of the Demicco risk models that replaces the mitotic count with the Ki-67 labeling index. We compared the three-variable and four-variable Demicco risk models with our modified risk models using Kaplan–Meier curves based on data for 43 patients with SFT.

Results

We found a significant difference in metastasis-free survival when patients were classified into low-risk and intermediate/high-risk groups using the three-variable (P = 0.022) and four-variable (P = 0.046) Demicco models. There was also a significant difference in metastasis-free survival between the low-risk and intermediate/high-risk groups when the modified three-variable (P = 0.006) and four-variable (P = 0.022) models were used.

Conclusion

Modified risk models that include the Ki-67 labeling index are effective for prediction of the prognosis in patients with SFT.

CT Radiomics for the Preoperative Prediction of Ki67 Index in Gastrointestinal Stromal Tumors: A Multi-Center Study

Purpose

This study established and verified a radiomics model for the preoperative prediction of the Ki67 index of gastrointestinal stromal tumors (GISTs).

Materials and Methods

A total of 344 patients with GISTs from three hospitals were divided into a training set and an external validation set. The tumor region of interest was delineated based on enhanced computed-tomography (CT) images to extract radiomic features. The Boruta algorithm was used for dimensionality reduction of the features, and the random forest algorithm was used to construct the model for radiomics prediction of the Ki67 index. The receiver operating characteristic (ROC) curve was used to evaluate the model’s performance and generalization ability.

Results

After dimensionality reduction, a feature subset having 21 radiomics features was generated. The generated radiomics model had an the area under curve (AUC) value of 0.835 (95% confidence interval(CI): 0.761–0.908) in the training set and 0.784 (95% CI: 0.691–0.874) in the external validation cohort.

Conclusion

The radiomics model of this study had the potential to predict the Ki67 index of GISTs preoperatively.

Ki-67 immunoexpression and radiological assessment of necrosis improves accuracy of conventional and modified core biopsy systems in predicting the final grade assigned to adult-soft tissue sarcomas. An international collaborative study

Based on the French Federation Nationale des Centers de Lutte Contre le Cancer (FNCLCC) grading system, this study assesses the accuracy of conventional and modified core biopsy (CB) systems in predicting the final grade (low vs high) assigned to the resected specimen. Substituting Ki-67 immunoexpression for mitotic count, and radiological for histological assessment of necrosis, we used two modified FNCLCC CB grading systems: (1) Ki-67 immunoexpression alone, and (2) Ki-67 plus radiological assessment of necrosis. We graded 199 soft tissue sarcomas (STS) from nine centers, and compared the results for the conventional (obtained from local histopathology reports) and modified CB systems with the final FNCLCC grading of the corresponding resected specimens. Due to insufficient sample quality or lack of available radiologic data, five cases were not evaluated for Ki67 or radiological assessment of necrosis. The conventional FNCLCC CB grading system accurately identified 109 of the 130 high-grade cases (83.8%). The CB grading matched the final FNCLCC grading (low vs high) in 175 (87.9%) of the 199 resected tumors; overestimating the final grade in three cases and underestimating in 21 cases. Modified system 1 (Ki-67) accurately identified 117 of the 130 high-grade cases (90.0%). The CB grading matched the final FNCLCC grading (low vs high) in 175 (89.7%) of the 195 evaluated cases; overestimating seven and underestimating 13 cases. Modified system 2 (Ki-67 plus radiological necrosis) accurately identified 120 of the 130 high-grade cases (92.3%). This last matched the final FNCLCC grading (low vs high) in 177 (91.2%) of the 194 evaluated cases; overestimating seven and underestimating 10 cases. Modified system 2 obtained highest area under ROC curves, although not statistically significant. Underestimated CB grades did not correlate with histological subtypes, although many of the discrepant cases were myxoid tumors (myxofibrosarcomas or myxoid liposarcomas), leiomyosarcomas or undifferentiated pleomorphic/spindle cell sarcomas. Using modified FNCLCC CB grading systems to replace conventional mitotic count and histologic assessment of necrosis may improve the distinction between low and high-grade STS on CB. Our study confirms that classifying grade 1 as low grade and grades 2 and 3 as high grade improves correlation between CB and final grade by up to 21%, irrespective of CB system used. A higher than expected Ki-67 score in a low-grade sarcoma diagnosed on CB should raise concern that a higher-grade component may not have been sampled. Furthermore, correlation of all clinicopathological and radiological findings at multidisciplinary meetings is essential to assess the histological grade on CB as accurately as possible.

Preoperative imaging of gastric GISTs underestimates pathologic tumor size: A retrospective, single institution analysis

BACKGROUND

How well imaging size agrees with pathologic size of gastric gastrointestinal stromal tumors (GISTs) is unknown. GIST risk stratification is based on pathologic size, location, and mitotic rate. To inform decision making, the size discrepancy between imaging and pathology for gastric GISTs was investigated.

METHODS

Imaging and pathology reports were reviewed for 113 patients. Bland-Altman analyses and intraclass correlation (ICC) assessed agreement of imaging and pathology. Changes in clinical risk category due to size discrepancy were identified.

RESULTS

Computed tomography (CT) (n = 110) and endoscopic ultrasound (EUS) (n = 50) underestimated pathologic size for gastric GISTs by 0.42 cm, 95% confidence interval (CI): (0.11, 0.73), p = 0.008 and 0.54 cm, 95% CI: (0.25, 0.82), p < 0.001, respectively. ICCs were 0.94 and 0.88 for CT and EUS, respectively. For GISTs ≤ 3 cm, size underestimation was 0.24 cm for CT (n = 28), 95% CI: (0.01, 0.47), p = 0.039 and 0.56 cm for EUS (n = 26), 95% CI: (0.27, 0.84), p < 0.0001. ICCs were 0.72 and 0.55 for CT and EUS, respectively. Spearman's correlation was ≥0.84 for all groups. For GISTs ≤ 3 cm, 6/28 (21.4% p = 0.01) on CT and 7/26 (26.9% p = 0.005) on EUS upgraded risk category using pathologic size versus imaging size. No GISTs ≤ 3 cm downgraded risk categories. Size underestimation persisted for GISTs ≤ 2 cm on EUS (0.39 cm, 95% CI: [0.06, 0.72], p = 0.02, post hoc analysis).

CONCLUSION

Imaging, particularly EUS, underestimates gastric GIST size. Caution should be exercised using imaging alone to risk-stratify gastric GISTs, and to decide between surveillance versus surgery.

Ki-67 Proliferation Index Assessment in Gastroenteropancreatic Neuroendocrine Tumors by Digital Image Analysis With Stringent Case and Hotspot Level Concordance Requirements

OBJECTIVES

The Ki-67 proliferation index is integral to gastroenteropancreatic neuroendocrine tumor (GEP-NET) assessment. Automated Ki-67 measurement would aid clinical workflows, but adoption has lagged owing to concerns of nonequivalency. We sought to address this concern by comparing 2 digital image analysis (DIA) platforms to manual counting with same-case/different-hotspot and same-hotspot/different-methodology concordance assessment.

METHODS

We assembled a cohort of GEP-NETs (n = 20) from 16 patients. Two sets of Ki-67 hotspots were manually counted by three observers and by two DIA platforms, QuantCenter and HALO. Concordance between methods and observers was assessed using intraclass correlation coefficient (ICC) measures. For each comparison pair, the number of cases within ±0.2xKi-67 of its comparator was assessed.

RESULTS

DIA Ki-67 showed excellent correlation with manual counting, and ICC was excellent in both within-hotspot and case-level assessments. In expert-vs-DIA, DIA-vs-DIA, or expert-vs-expert comparisons, the best-performing was DIA Ki-67 by QuantCenter, which showed 65% cases within ±0.2xKi-67 of manual counting.

CONCLUSIONS

Ki-67 measurement by DIA is highly correlated with expert-assessed values. However, close concordance by strict criteria (>80% within ±0.2xKi-67) is not seen with DIA-vs-expert or expert-vs-expert comparisons. The results show analytic noninferiority and support widespread adoption of carefully optimized and validated DIA Ki-67.

Correlation Between Digital and Manual Determinations of Ki-67/MIB-1 Proliferative Indices in Human Meningiomas

Objective. Proliferative activity in tumor tissues is assessed as the percentage of Ki-67/MIB-1-positive cells, or the proliferative index (PI). The PI is routinely assessed manually. However, the subjectivity of manual assessments might result in poor reproducibility. We hypothesized that digital assessments might reduce the error. Method. In our study, we assessed Ki-67/MIB-1 PIs, both manually and digitally, with tissue microarrays constructed from 141 human meningioma samples. Spearman-rank correlation and κ statistics were applied for correlation and agreement analyses, respectively. Mann-Whitney U tests were used to compare MIB-1 PIs between groups. Prognostic ability was assessed with Kaplan-Meier and Cox regression analyses. Results. We found a significant, high correlation (Spearman ρ = 0.832, P < .01) and moderate agreement (κ coefficient = 0.617, observed agreement = 80.9%) between the 2 methods. Both methods found significantly different Ki-67/MIB-1 PIs for different World Health Organization grades (P < .05). Neither method showed significant prognostic value. Conclusion. Digital determinations of Ki-67/MIB-1 PIs in human meningiomas are feasible for the daily routine.

Surgical treatment of gastrointestinal stromal tumours combined with imatinib treatment: a retrospective cohort analysis

BACKGROUND

Targeted therapies changed the treatment concepts of gastrointestinal stromal tumours significantly. As only possibility to cure surgical resection is the cornerstone of therapy. Thus it is necessary to find out which patients will benefit most regarding modality (neo- or adjuvant) and duration of chemotherapy.

METHODS

In a retrospective cohort analysis the medical records of all consecutive patients treated in the department of general and visceral surgery of the university hospital Tübingen between 2004 and 2015 were investigated. Recurrence and survival outcomes were calculated using the Kaplan-Meier method.

RESULTS

Tumor location of GIST was gastric in 32, small bowel in 14, rectum in 3 and extraintestinal in 3 patients. Median tumor size was 46 mm. Median mitotic index was 4 per 50 hpf. Resection was achieved R0 in 46 patients, R1 in 4 patients and R2 in 2 patients. Mean overall survival was 58.9 months (range, 46-73 months). Mean recurrence free survival was 45.6 months (range, 36-57 months). Mean overall survival was 58.9 months (range, 46-73 months). Risk factors for recurrence were tumor location and high mitotic index Ki-67.

CONCLUSIONS

The prognosis of GIST after surgical resection is favourable. Survival prognosis are excellent. Analysis of KI-67 mitotic index predicted best oncologic outcome.

Tumor Digital Masking Allows Precise Patient Triaging: A Study Based on Ki-67 Scoring in Gastrointestinal Stromal Tumors

BACKGROUND

Technological advances constantly provide cutting-edge tools that enhance the progress of diagnostic capabilities. Gastrointestinal stromal tumors belong to a family of mesenchymal tumors where patient triaging is still based on traditional criteria such as mitotic count, tumor size, and tumor location. Limitations of the human eye and randomness in choice of area for mitotic figure counting compel us to seek more objective solutions such as digital image analysis. Presently, the labelling of proliferative activity is becoming a routine task amidst many cancers. The purpose of the present study was to compare the traditional method of prediction based on mitotic ratio with digital image analysis of cell cycle-dependent proteins.

METHODS

Fifty-seven eligible cases were enrolled. Furthermore, a digital analysis of previously performed whole tissue section immunohistochemical assays was executed. Digital labelling covered both hotspots and not-hotspots equally.

RESULTS

We noted a significant diversity of proliferative activities, and consequently, the results pointed to 6.5% of Ki-67, counted in hotspots, as the optimal cut-off for low-high-grade GIST. ROC analysis (AUC = 0.913; 95% CI: 0.828-0.997, p < 0.00001) and odds ratio (OR = 40.0, 95% CI: 6.7-237.3, p < 0.0001) pointed to Ki-67 16% as the cut-off for very high-grade (groups 5-6) cases. With help of a tumor digital map, we revealed possible errors resulting from a wrong choice of field for analysis. We confirmed that Ki-67 scores are in line with the level of intracellular metabolism that could be used as the additional biomarker.

CONCLUSIONS

Tumor digital masking is very promising solution for repeatable and objective labelling. Software adjustments of nuclear shape, outlines, size, etc. are helpful to omit other Ki-67-positive cells especially small lymphocytes. Our results pointed to Ki-67 as a good biomarker in GIST, but concurrently, we noted significant differences in used digital approaches which could lead to unequivocal results.