Comparison of nomograms and predictive tools in gastrointestinal stromal tumor

Introduction—heterogeneity of gastrointestinal stromal tumors (GISTs)

GISTs have significant heterogeneity in tumor biology with varying rates of survival (5-year overall survival: 59–81.3%) (1,2) and recurrence (3.8–37.5%) (3,4). Given such heterogeneity, multiple studies have attempted to identify factors associated with increased risks of recurrence and worse survival. These studies may be grouped largely by: (I) patient-related and pathological variables; (II) molecular/genomic variables; (III) radiographic variables.

Patient-related and pathological variables

Previous studies identify tumor location and sex as important predictors of recurrence. In a pooled analysis of 2,560 patients, Joensuu et al. demonstrated that gastric GISTs were associated with better recurrence-free survival (RFS) than intestinal GISTs [hazard ratio (HR) 0.49, 95% confidence interval (CI): 0.35–0.67] (5). GISTs located outside the gastrointestinal tract had the worst overall survival (HR 2.13, 95% CI: 1.44–3.15). In addition, female sex was associated with improved RFS (HR 0.90, 95% CI: 0.82–0.99). Similar findings were reported by Rutkowski et al. where patients with gastric GISTs were associated with better RFS compared to other anatomic sites (HR 0.62, 95% CI: 0.41–0.95) and male patients had an associated worse RFS (HR 1.78, 95% CI: 1.21–2.61) (6). DeMatteo et al. previously described several additional pathological risk factors, such as higher mitotic rate, larger tumor size, and small bowel primaries, which had the highest risk of recurrence (7), while Joensuu et al.’s further identified tumor rupture as an independent risk factor for recurrence (5). The clinical implications of tumor rupture were further explored using data from the SSGXVIII/AIO clinical trial, showing that patients with tumor rupture had a higher rate of recurrence than their unruptured counterparts (8).

Inflammatory markers and the gut microbiome are emerging as prognostic markers associated with significant physiological stress and are under investigation in GISTs (9,10). Sarhadi et al. found that Enterobacteriaceae abundance was associated with more aggressive gastric tumors (11), while Li et al. further expanded on this data by showing the presence of Shewanella algae within GISTs as a potential risk factor for tumor invasion (12). Zhao et al. demonstrated an association between RFS and inflammatory biomarkers in patients with GIST (13). Research to define the prognostic role of inflammatory markers and gut microbiome are emerging, important questions for the future.

Molecular/genomic variables

The rise in tumor molecular testing led to the identification of specific mutations that helped to explain some of the heterogeneity in outcomes and sensitivity to certain adjuvant treatment types, doses, and duration (14). For example, GISTs are divided into imatinib-sensitive and non-sensitive tumors. The imatinib-sensitive GISTs are those tumors with mutations in KIT exon 11, KIT exon 9, and PDGFRA (non-D842V). In the landmark SSGXVIII/AIO clinical trial, the study showed that 3 years of imatinib treatment led to improved RFS and overall survival compared to 1 year of adjuvant imatinib. The importance of molecular data in understanding heterogeneous outcomes in GIST was further demonstrated in this trial by showing that patients with PDGFRA and KIT exon 11 mutations had a favorable RFS compared to KIT or PDGFRA mutation wild-type patients, while those with KIT exon 9 did not demonstrate a statistically significant benefit (15).

Imatinib-non-sensitive GIST mutations may be divided into those with PDGFRA D842V mutation, BRAF V600E mutation, SDH deficiency, and NTRK rearrangements. The NAVIGATOR and CS3007-001 trials are actively investigating the role of avapritinib in the treatment of patients with the PDGFRA D842V mutation (16). Despite their insensitivity to imatinib, D842V mutants are known to have a low mitotic count and favorable prognosis, which is highlighted in the SSGXVIII/AIO clinical trial with an overall survival rate of 90%, independent of imatinib therapy (17). Sasa et al. investigated the impact of NTRK2 gene expression in GIST and found that these expressions are usually present in the duodenum and small intestine, and these patients experience a shorter disease-free survival (18). Miranda et al., utilizing the Ticino Cancer Center Registry database, demonstrated that 2% of patients with GISTs carried a BRAF mutation and demonstrated that imatinib was unable to inhibit signaling from RAS-RAF mutations (19). Identification of molecular variables has led to significantly improved capacity in explaining some of the heterogeneity in GIST outcomes and response to therapy.

Radiographic variables

Investigators have also sought to identify radiologic features of patients with GISTs who are at a higher risk of recurrence, particularly after surgical resection (20-22). Chen et al. conducted a retrospective analysis of the preoperative CT findings in 132 patients with gastric GIST (22). Features included tumor size and volume, calcification, ulceration, serosal invasion, mean attenuation, mean enhancement, and growth patterns. Serosal invasion and high enhancement were independently associated with a worse disease-free survival in the multivariate model. Grazzini et al. further supported the use of certain radiological features as risk factors of GIST recurrence (21). In a retrospective analysis of 54 patients, preoperative CT findings of tumor necrosis and enlarged feeding or draining vessels were identified as the risk factors for recurrence. Furthermore, retrospective data suggests that CT findings correlate with tumor grade and 5-year mortality (23). Together, these findings suggest that certain radiological features may serve as surrogates for more aggressive tumor biology.

CT findings are often subjective, prompting growing interest in artificial intelligence for more objective assessment. Chen et al. showed that a deep learning radiomics model improved prediction of RFS in patients with GISTs (24). In an external validation cohort, the model achieved an area under curve (AUC) of 0.887 (95% CI: 0.816–0.960), outperforming the clinicopathologic nomogram (AUC 0.772, 95% CI: 0.679–0.865) and traditional risk stratification systems, including the modified National Institutes of Health (NIH) (AUC 0.754, 95% CI: 0.667–0.841) and Armed Forces Institute of Pathology (AFIP) (AUC 0.739, 95% CI: 0.643–0.835). A nomogram incorporating tumor size, location, mitotic index, and the deep learning radiomics model further improved prediction of 5-year disease-free survival (AUC 0.923, 95% CI: 0.812–0.999). These findings highlight the potential for AI to complement radiologic interpretation and strengthen prognostic nomograms.

Newly gained insights from article of interest

In a large multimodal model, Yang et al. have elucidated factors associated with recurrence. Multivariate logistic regression analysis demonstrated that mitotic count, Ki-67 labeling index, GIST site and size, mass enhancement pattern, tumor density in the plain phase, cystic change, and adjacent organ invasion were significant risk factors for recurrence (25). Cox analysis further demonstrated that mitotic count, Ki-67 labeling index, tumor site and size, mass enhancement pattern, necrosis content, cystic change, and peritumoral fat infiltration were significant predictors of overall survival. Further studies using machine learning and incorporating molecular data may further improve risk prediction.

Prognostication tools for GIST

Initial efforts—risk stratification models

Using the information from studies mentioned above, various prognostic tools have been proposed for patients with GISTs. In 2002, Fletcher et al. established the NIH stratification system (26). The model includes tumor size and mitotic index to stratify patients into risk categories (very low-risk, low-risk, intermediate-risk, and high-risk). In 2006, Miettinen and Lasota published the AFIP risk stratification criteria, which considered pathological factors such as mitotic index, tumor site, and tumor size (27). In 2008, Joensuu et al. proposed a modification to the NIH stratification system that accounted for tumor rupture while also stratifying tumor size and mitotic count by primary site (28). Despite the establishment of these stratification tools, studies have demonstrated variable predictive performance, indicating the need to refine the inclusion of risk factors that predict outcomes in patients at risk for recurrence (29-31). These findings highlight some key limitations of categorical stratification systems. Therefore, the use of nomograms has progressed, as these statistical models incorporate multiple prognostic variables in a weighted and continuous manner.

Rise of nomograms

Nomograms are predictive statistical models that can generate an individualized probability of a clinical event, such as recurrence, using specific clinical, pathologic, or radiographic variables (32). In addition, nomograms tend to be user-friendly tools that often demonstrate improved accuracy over conventional staging models (33).

Gold and colleagues at the Memorial Sloan Kettering Cancer Center (MSKCC) previously developed and validated a nomogram for evaluating RFS after R0 resections in GIST (34). The nomogram was modeled using data from 127 patients with information on tumor size, location, and mitotic index. The model significantly outperformed the concordance probabilities of the standard and modified NIH staging systems. This nomogram was further externally validated by Chok et al. using a single-institutional cohort of patients, with a comparison to the NIH, AFIP, and modified NIH systems (35). Chen et al. compared performance for predicting recurrence using multi-center data and used receiver operating curves to compare the accuracy between the MSKCC nomogram, NIH, and AFIP staging system (31). In their study, the AFIP criteria had the highest accuracy in recurrence prediction (AUC AFIP: 0.75 vs. MSKCC 2-year: 0.74). Therefore, some have expanded on the MSKCC nomogram to include other variables to expand the predictive capabilities of GIST recurrence.

Next generation of prognostic models

Newer predictive tools integrate clinical, pathologic, molecular, and radiologic data. Bertsimas et al. developed an artificial intelligence prediction model using demographic and pathologic data that demonstrated superior discrimination and decision curve analysis compared with the MSKCC nomogram (concordance index: 0.805, 95% CI: 0.803–0.808 vs. 0.788, 95% CI: 0.786–0.791, respectively) (36). While Bertsimas et al. did not include radiological data, other studies have incorporated radiology in their models (37-39). Yingzheng et al. constructed a nomogram from the CT features in 250 patients. Incorporating radiologic features such as tumor size, necrosis, and morphological characteristics helped predict patients who may have a high or low mitotic index. Other efforts have been made to include both radiologic and pathologic factors to predict GIST recurrence (40-42). Xiao et al. previously published a multimodal deep learning predictive model using CT features, pathology slides, and clinical data (42). In their external validation, the multimodal model (C-index of 0.864) outperformed the standalone clinical (C-index of 0.767), pathologic (C-index of 0.802), and radiologic models (C-index of 0.777). Together these findings provide validation for Yang et al.’s hypothesis regarding the synergy between radiological and pathologic features for model construction.

More recently, these prognostic models have incorporated genomic data. Dermawan et al. developed a machine learning model combining clinical and genomic data to develop a three-tier risk stratification system (43). For example, gastric GISTs were high-risk if they had a chr1p or SDHB mutation and intermediate risk if a chr14q mutation was present or a KIT exon 11 mutation was absent. Small bowel GISTs were at high risk if they had a MYC, CDKN2A or RB1 mutation, while patients with a chr1p mutation or chr5q alteration were classified as intermediate risk. The model was able to predict a significant difference in RFS between the low- and high-risk groups (P<0.01) and intermediate- and high-risk groups (P<0.01). In comparison to the traditional risk stratification systems (AFIP, NIH, Joensuu et al. system), this study demonstrated that 10–28% of patients who were traditionally at low risk were upgraded to high genomic risk; 10–13% of patients traditionally high-risk were downgraded to low genomic risk; and 54–57% were downgraded to moderate genomic risk.

Newly gained insights from article of interest

Yang et al. designed a nomogram that integrates radiological findings from computed tomography (CT) images, postoperative pathology, and patient-related clinical data to predict recurrence and long-term prognosis in patients with resected GISTs. The authors compared the performance of the nomogram model with that of the modified NIH classification system. The combined model showed significantly improved performance over the modified NIH model in the validation (AUC: 0.810 vs. 0.739, P=0.005) and external test (AUC: 0.953 vs. 0.894, P=0.043). In contrast, there were no significant differences between the CT-based nomogram and modified NIH system in the validation (AUC: 0.786 vs. 0.739, P=0.413) and external cohorts (AUC: 0.917 vs. 0.894, P=0.552). These results highlight the synergistic predictive power of models that combine pathology, clinical, and radiologic data in nomograms. Furthermore, the authors created “high-recurrence” and “low-recurrence” risk groups through their combined prognostic nomogram. The nomogram was able to predict prognosis well: high recurrence group demonstrating RFS rates at 1-year 98.9%; 3-year 88.6%; 5-year 78.9%; 7-year 67.8% vs. low recurrence group RFS at 1-year 100%; 3-year 100%; 5-year 95.9%; 7-year 81% in the validation groups. These findings highlight the value of Yang et al.’s nomogram in the differential RFS of high-risk patients, where previous models had difficulty (29,30).

Practical uses of nomograms for individualized approach

Nomograms give patients and clinicians the ability to translate risk factors and clinical data into a result that informs them of various prognostic information such as RFS and overall survival. Additionally, some nomograms incorporate actionable therapies to inform patients about the potential benefits of certain treatments to their outcomes (44). Therefore, nomograms move beyond traditional risk stratification by providing personalized, therapy-specific prognostic information that can help guide clinical decision-making in patients.

Recurrence after curative-intent resection remains a significant problem in patients with GIST (45). Approximately 80% of tumors harbor KIT mutations, making imatinib an effective targeted therapy in most patients. The ACOSOG Z9001 trial demonstrated superior RFS after adjuvant imatinib (HR 0.6, 95% CI: 0.43–0.75) (46,47). Since the report of these landmark trials, subsequent clinical trials have sought to better define the benefit of adjuvant imatinib among patients meeting refined high-risk criteria. The PERSIST-5 trial investigated the benefit of adjuvant imatinib in patients at a high risk of recurrence after surgical resection of GIST defined as those with a GIST at any site measuring ≥2 cm with ≥5 mitoses per 50 high power field (HPF) or non-gastric primary GIST measuring ≥5 cm (48). At the 5-year follow-up, the patients with imatinib-sensitive tumors on imatinib throughout 5 years remained disease-free. To determine the optimal duration of imatinib therapy, the SSGXVII/AIO trial used the NIH consensus criteria, which defines high-risk as tumors >10 cm, >10 mitoses per 50 HPF, or tumor size >5.0 cm with >5 mitoses per 50 HPF, or tumor rupture (49). Together, these trials highlight the importance of identifying high-risk patients, as risk stratification directly influences both the prognostic performance of nomograms and individualized selection of adjuvant imatinib.

Newly gained insights from article of interest

In their subgroup analysis, Yang et al. evaluated the prognostic differentiation of their model between non-adjuvant therapy and adjuvant therapy groups. In the non-adjuvant therapy subgroup, the model demonstrated a validation C-index of 0.812 (95% CI: 0.748–0.875) and an external test C-index of 0.944 (95% CI: 0.893–0.994). In contrast, in the adjuvant therapy subgroup, the prognostic model had a validation C-index of 0.755 (95% CI: 0.683–0.879) and an external test C-index of 0.755 (95% CI: 0.636–0.873). These results suggest that the model strongly differentiated patients who did not receive adjuvant therapy, whereas prognostic separation may be attenuated in patients who received adjuvant therapy. Ultimately, the nomogram showed strong discriminative ability across cohorts and may serve as a useful tool for predicting prognosis in patients with or without adjuvant imatinib therapy.

Conclusions

Yang et al. developed an integrative nomogram incorporating radiological, pathological, and clinical factors to predict recurrence and overall survival in patients with GIST, outperforming CT-only models and modified NIH criteria. These findings support multi-perspective nomograms, which may further improve by adding molecular/genomics data. Further international external validation of this model has the potential to provide a superior stratification of patients with GIST who would benefit from adjuvant therapy to better personalize the care of such a heterogeneous group of patients.

Acknowledgments

None.

Footnote

Provenance and Peer Review: This article was commissioned by the editorial office, Journal of Gastrointestinal Oncology. The article has undergone external peer review.

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-1-0192/prf

Funding: None.

Conflicts of Interest: Both authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-1-0192/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Li J, Huang Z, Zhou H, et al. Survival outcomes and prognostic factors of advanced gastrointestinal stromal tumors: in the era of multiple tyrosine kinase inhibitors. J Gastrointest Oncol 2024;15:931-45. [Crossref] [PubMed]
2. Zhu H, Yang G, Ma Y, et al. Update of epidemiology, survival and initial treatment in patients with gastrointestinal stromal tumour in the USA: a retrospective study based on SEER database. BMJ Open 2023;13:e072945. [Crossref] [PubMed]
3. Guérin A, Sasane M, Keir CH, et al. Physician Underestimation of the Risk of Gastrointestinal Stromal Tumor Recurrence After Resection. JAMA Oncol 2015;1:797-805. [Crossref] [PubMed]
4. Bleckman RF, Roets E, IJzerman NS, et al. Local recurrence in primary localised resected gastrointestinal stromal tumours: A registry observational national cohort study including 912 patients. Eur J Cancer 2023;186:113-21. [Crossref] [PubMed]
5. Joensuu H, Vehtari A, Riihimäki J, et al. Risk of recurrence of gastrointestinal stromal tumour after surgery: an analysis of pooled population-based cohorts. Lancet Oncol 2012;13:265-74. [Crossref] [PubMed]
6. Rutkowski P, Nowecki ZI, Michej W, et al. Risk criteria and prognostic factors for predicting recurrences after resection of primary gastrointestinal stromal tumor. Ann Surg Oncol 2007;14:2018-27. [Crossref] [PubMed]
7. Dematteo RP, Gold JS, Saran L, et al. Tumor mitotic rate, size, and location independently predict recurrence after resection of primary gastrointestinal stromal tumor (GIST). Cancer 2008;112:608-15. [Crossref] [PubMed]
8. Joensuu H, Reichardt A, Eriksson M, et al. Survival of patients with ruptured gastrointestinal stromal tumour treated with adjuvant imatinib in a randomised trial. Br J Cancer 2024;131:299-304. [Crossref] [PubMed]
9. Haghi SE, Khanzadeh M, Sarejloo S, et al. Systematic review of the significance of neutrophil to lymphocyte ratio in anastomotic leak after gastrointestinal surgeries. BMC Surg 2024;24:15. [Crossref] [PubMed]
10. Panther EJ, Dodd W, Clark A, et al. Gastrointestinal Microbiome and Neurologic Injury. Biomedicines 2022;10:500. [Crossref] [PubMed]
11. Sarhadi V, Mathew B, Kokkola A, et al. Gut microbiota of patients with different subtypes of gastric cancer and gastrointestinal stromal tumors. Gut Pathog 2021;13:11. [Crossref] [PubMed]
12. Li Y, Zhang R, Fu C, et al. Intratumoral microbiome promotes liver metastasis and dampens adjuvant imatinib treatment in gastrointestinal stromal tumor. Cancer Lett 2024;601:217149. [Crossref] [PubMed]
13. Zhao JL, Wang MY, Lv YZ, et al. Prognostic value of inflammatory markers in predicting recurrence-free survival in gastrointestinal stromal tumor patients: A nomogram-based approach. World J Gastrointest Oncol 2025;17:94956. [Crossref] [PubMed]
14. Catalano F, Cremante M, Dalmasso B, et al. Molecular Tailored Therapeutic Options for Advanced Gastrointestinal Stromal Tumors (GISTs): Current Practice and Future Perspectives. Cancers (Basel) 2023;15:2074. [Crossref] [PubMed]
15. Joensuu H, Wardelmann E, Sihto H, et al. Effect of KIT and PDGFRA Mutations on Survival in Patients With Gastrointestinal Stromal Tumors Treated With Adjuvant Imatinib: An Exploratory Analysis of a Randomized Clinical Trial. JAMA Oncol 2017;3:602-9. [Crossref] [PubMed]
16. Heinrich MC, Zhang X, Jones RL, et al. Clinical Benefit of Avapritinib in KIT-Mutant Gastrointestinal Stromal Tumors: A Post Hoc Analysis of the Phase I NAVIGATOR and Phase I/II CS3007-001 Studies. Clin Cancer Res 2024;30:719-28. [Crossref] [PubMed]
17. Joensuu H, Wardelmann E, Eriksson M, et al. KIT and PDGFRA Mutations and Survival of Gastrointestinal Stromal Tumor Patients Treated with Adjuvant Imatinib in a Randomized Trial. Clin Cancer Res 2023;29:3313-9. [Crossref] [PubMed]
18. Sasa K, Son R, Oguchi A, et al. NTRK2 expression in gastrointestinal stromal tumors with a special emphasis on the clinicopathological and prognostic impacts. Sci Rep 2024;14:768. [Crossref] [PubMed]
19. Miranda C, Nucifora M, Molinari F, et al. KRAS and BRAF mutations predict primary resistance to imatinib in gastrointestinal stromal tumors. Clin Cancer Res 2012;18:1769-76. [Crossref] [PubMed]
20. Park JH, Kang BK, Lee HL, et al. Correlation between Three-Dimensional Volume and Malignant Potential of Gastrointestinal Stromal Tumors (GISTs). J Clin Med 2020;9:2763. [Crossref] [PubMed]
21. Grazzini G, Guerri S, Cozzi D, et al. Gastrointestinal stromal tumors: relationship between preoperative CT features and pathologic risk stratification. Tumori 2021;107:556-63. [Crossref] [PubMed]
22. Chen XS, Yuan W, Xu ZH, et al. Prognostic value of preoperative CT features for disease-free survival in patients with primary gastric gastrointestinal stromal tumors after resection. Abdom Radiol (NY) 2023;48:494-501. [Crossref] [PubMed]
23. Tateishi U, Hasegawa T, Satake M, et al. Gastrointestinal stromal tumor. Correlation of computed tomography findings with tumor grade and mortality. J Comput Assist Tomogr 2003;27:792-8. [Crossref] [PubMed]
24. Chen T, Liu S, Li Y, et al. Developed and validated a prognostic nomogram for recurrence-free survival after complete surgical resection of local primary gastrointestinal stromal tumors based on deep learning. EBioMedicine 2019;39:272-9. [Crossref] [PubMed]
25. Yang CW, Yu X, Che F, et al. Development and validation of CT-based nomograms for predicting recurrence and prognosis in gastrointestinal stromal tumor. Abdom Radiol (NY) 2026;51:2292-303. [Crossref] [PubMed]
26. Fletcher CD, Berman JJ, Corless C, et al. Diagnosis of gastrointestinal stromal tumors: a consensus approach. Int J Surg Pathol 2002;10:81-9. [Crossref] [PubMed]
27. Miettinen M, Lasota J. Gastrointestinal stromal tumors: pathology and prognosis at different sites. Semin Diagn Pathol 2006;23:70-83. [Crossref] [PubMed]
28. Joensuu H. Risk stratification of patients diagnosed with gastrointestinal stromal tumor. Hum Pathol 2008;39:1411-9. [Crossref] [PubMed]
29. Schmieder M, Henne-Bruns D, Mayer B, et al. Comparison of Different Risk Classification Systems in 558 Patients with Gastrointestinal Stromal Tumors after R0-Resection. Front Pharmacol 2016;7:504. [Crossref] [PubMed]
30. Goh BK, Chow PK, Yap WM, et al. Which is the optimal risk stratification system for surgically treated localized primary GIST? Comparison of three contemporary prognostic criteria in 171 tumors and a proposal for a modified Armed Forces Institute of Pathology risk criteria. Ann Surg Oncol 2008;15:2153-63. [Crossref] [PubMed]
31. Chen T, Ye LY, Feng XY, et al. Performance of risk stratification systems for gastrointestinal stromal tumors: A multicenter study. World J Gastroenterol 2019;25:1238-47. [Crossref] [PubMed]
32. Iasonos A, Schrag D, Raj GV, et al. How to build and interpret a nomogram for cancer prognosis. J Clin Oncol 2008;26:1364-70. [Crossref] [PubMed]
33. Balachandran VP, Gonen M, Smith JJ, et al. Nomograms in oncology: more than meets the eye. Lancet Oncol 2015;16:e173-80. [Crossref] [PubMed]
34. Gold JS, Gönen M, Gutiérrez A, et al. Development and validation of a prognostic nomogram for recurrence-free survival after complete surgical resection of localised primary gastrointestinal stromal tumour: a retrospective analysis. Lancet Oncol 2009;10:1045-52. [Crossref] [PubMed]
35. Chok AY, Goh BK, Koh YX, et al. Validation of the MSKCC Gastrointestinal Stromal Tumor Nomogram and Comparison with Other Prognostication Systems: Single-Institution Experience with 289 Patients. Ann Surg Oncol 2015;22:3597-605. [Crossref] [PubMed]
36. Bertsimas D, Margonis GA, Tang S, et al. An interpretable AI model for recurrence prediction after surgery in gastrointestinal stromal tumour: an observational cohort study. EClinicalMedicine 2023;64:102200. [Crossref] [PubMed]
37. Ji X, Shang Y, Tan L, et al. Prediction of High-Risk Gastrointestinal Stromal Tumor Recurrence Based on Delta-CT Radiomics Modeling: A 3-Year Follow-up Study After Surgery. Clin Med Insights Oncol 2024;18:11795549241245698. [Crossref] [PubMed]
38. Yingzheng R, Linlin J, Yang Y, et al. Study on the predictive value of preoperative CT features for the mitotic index of GIST based on the nomogram. Sci Rep 2025;15:8627. [Crossref] [PubMed]
39. Jung H, Lee SM, Kim YC, et al. Gastrointestinal stromal tumours: Preoperative imaging features to predict recurrence after curative resection. Eur J Radiol 2022;149:110193. [Crossref] [PubMed]
40. Ran P, Tan T, Zhou H, et al. Nomogram for Predicting Recurrence-Free Survival of Primary Localized Gastrointestinal Stromal Tumor. J Pers Med 2023;13:498. [Crossref] [PubMed]
41. Zheng J, Liao Q, Chen X, et al. Development and validation of a computed tomography-based radiomics signature to predict “highest-risk” from patients with high-risk gastrointestinal stromal tumor. J Gastrointest Oncol 2024;15:125-33. [Crossref] [PubMed]
42. Xiao X, Han X, Sun Y, et al. Development and interpretation of a multimodal predictive model for prognosis of gastrointestinal stromal tumor. NPJ Precis Oncol 2024;8:157. [Crossref] [PubMed]
43. Dermawan JK, Kelly C, Gao Z, et al. Novel Genomic Risk Stratification Model for Primary Gastrointestinal Stromal Tumors (GIST) in the Adjuvant Therapy Era. Clin Cancer Res 2023;29:3974-85. [Crossref] [PubMed]
44. Albert JM, Liu DD, Shen Y, et al. Nomogram to predict the benefit of radiation for older patients with breast cancer treated with conservative surgery. J Clin Oncol 2012;30:2837-43. [Crossref] [PubMed]
45. DeMatteo RP, Lewis JJ, Leung D, et al. Two hundred gastrointestinal stromal tumors: recurrence patterns and prognostic factors for survival. Ann Surg 2000;231:51-8. [Crossref] [PubMed]
46. Corless CL, Ballman KV, Antonescu CR, et al. Pathologic and molecular features correlate with long-term outcome after adjuvant therapy of resected primary GI stromal tumor: the ACOSOG Z9001 trial. J Clin Oncol 2014;32:1563-70. [Crossref] [PubMed]
47. DeMatteo RP, Ballman KV, Antonescu CR, et al. Placebo-Controlled Randomized Trial of Adjuvant Imatinib Mesylate Following the Resection of Localized, Primary Gastrointestinal Stromal Tumor (GIST). Lancet 2009;373:1097-104. [Crossref] [PubMed]
48. Raut CP, Espat NJ, Maki RG, et al. Efficacy and Tolerability of 5-Year Adjuvant Imatinib Treatment for Patients With Resected Intermediate- or High-Risk Primary Gastrointestinal Stromal Tumor: The PERSIST-5 Clinical Trial. JAMA Oncol 2018;4:e184060. [Crossref] [PubMed]
49. Joensuu H, Eriksson M, Sundby Hall K, et al. One vs three years of adjuvant imatinib for operable gastrointestinal stromal tumor: a randomized trial. JAMA 2012;307:1265-72. [Crossref] [PubMed]