Page 38 - ARNM-3-2

Advances in Radiotherapy
& Nuclear Medicine Radiomics for gastric cancer

Table 3. Summary of studies on radiomics for predicting tumor prognosis and treatment response
Authors Year Image Purpose Sample Features Result
Giganti et al. 58 2017 CT To explore the relationship 56 (Single center) Radiomics A number of parameters were
between pre-operative CT significantly associated with the
texture analysis and OS in negative outcomes depending on the
patients with GC threshold
Wang et al. 59 2020 CT To evaluate the splenic 243 (Single center) Radiomics+clinical Splenic features extracted from
tissue characteristics to imaging technology can accurately
predict the prognosis of predict the long-term survival of
patients with GC patients with GC
Wang et al. 60 2020 CT To establish a prognosis 353 (Multi-center) Radiomics+clinical The radiomic nomogram
model guided by incorporated with radiomic
multi-detector CT signature, extramural vessel invasion,
clinical T stage, and clinical N stage
outperformed all the other models
(concordance index=0.720 and 0.727)
Li et al. 61 2019 CT To investigate the 181 (Single center) Radiomics+clinical The Harrell concordance index
prognostic significance of a nomogram combining
of radiomic features in radiomic signature and significant
patients with GC after clinicopathological risk factors was
radical resection 0.82
Li et al. 62 2018 CT To predict the pathological 47 (Single center) Radiomics The feature selection method
reaction of pre-operative adopted by a filter based on linear
chemotherapy for locally discriminant analysis+classifier
advanced GC of random forest achieved a
significantly prognostic performance
in the PP (AUC=0.722±0.108,
accuracy=0.793, sensitivity=0.636,
and specificity=0.889)
Sun et al. 63 2020 CT To predict the therapeutic 106 (Single center) Radiomics+clinical In the validation cohort, the rad-score
response to NAC and to demonstrated a good predicting
investigate its efficacy in performance in treatment response to
survival stratification the NAC (AUC=0.82)
Mazzei et al. 64 2021 CT To predict the response to 70 (Single center) Radiomics The AUC of all patients by logistic
NAC of GC regression was 0.763
Xu et al. 65 2021 CT To predict and early 292 (Single center) Radiomics The improved DR model with
detect the pathological averaging outcome scores of PR and
downstaging with NAC in DR models showed boosted results in
advanced GC two testing cohorts (AUC=0.961 and
AUC=0.921, respectively)
Chen et al. 66 2021 CT To predict the main 221 (Single center) Radiomics+clinical The final established model
pathological reactions of incorporates ADC differentiation
advanced GC to NAC and rad-scores. The model showed
satisfactory predictive accuracy with a
C-index of 0.763
Chen et al. 67 2021 CT To predict the DFS and OS 159 (Single center) Radiomics+clinical The combined Rad-clinical models
of patients with advanced showed improved performance in the
GC after NAC testing cohort, with C-indices of 0.810
and 0.710 for DFS and OS, respectively
Shin et al. 68 2021 CT To predict the 410 (Single center) Radiomics+clinical In internal and external validation,
recurrence-free survival of the AUC of the combined model was
locally advanced GC 0.719 and 0.651, respectively
Klaassen et al. 70 2018 CT To predict the 69 (Single center) Radiomics The random forest model for CT scan
chemotherapy response of lesions had an average training AUC
patients with esophageal of 0.87 and 0.79 for the validation set
GC to single hepatic
metastases
(Cont’d...)

Volume 3 Issue 2 (2025) 30 doi: 10.36922/arnm.8350