Tumor Compactness based on CT to predict prognosis after multimodal treatment for esophageal squamous cell carcinoma

Abstract

We aimed to establish a risk model using computed tomography-based compactness to predict overall survival (OS) and progression-free survival (PFS) after multimodal treatment for esophageal squamous cell carcinoma (ESCC). We extracted pre-treatment computed tomography-based tumor data (volume, surface area, and compactness) for 512 cases of ESCC that were treated at 3 centers. A risk model based on compactness was trained using Cox regression analyses of data from 83 cases, and then the model was validated using two independent cohorts (98 patients and 283 patients). The largest cohort (283 patients) was then evaluated using the risk model to predict response to radiotherapy with or without chemotherapy. In the three datasets, the pre-treatment compactness risk model provided good accuracy for predicting OS (P = 0.012, P = 0.022, and P = 0.003) and PFS (P < 0.001, P = 0.003, and P = 0.005). Patients in the low-risk group did not experience a significant OS benefit from concurrent chemoradiotherapy (P = 0.099). Furthermore, after preoperative concurrent chemoradiotherapy, the OS outcomes were similar among patients in the low-risk group who did and did not achieve a pathological complete response (P = 0.127). Tumor compactness was correlated with clinical T stage but was more accurate for predicting prognosis after treatment for ESCC, based on higher C-index values in all three datasets. The compactness-based risk model was effective for predicting OS and PFS after multimodal treatment for ESCC. Therefore, it may be useful for guiding personalized treatment.

A FDG-PET radiomics signature detects esophageal squamous cell carcinoma patients who do not benefit from chemoradiation

“…Nevertheless for potential non-responders to CRT it would be pivotal to identify these patients prior to therapy. The prognostic model may be further improved by the implementation of CT derived parameters, as a recent multicenter study showed promising results regarding risk stratification based on CT data 23 .…”

Methodological quality of machine learning-based quantitative imaging analysis studies in esophageal cancer: a systematic review of clinical outcome prediction after concurrent chemoradiotherapy

“…2008–2011 CT Inst 1: 120 kVp, 406 mAs, 3–5 mm Inst 2: 120 kVp, 150 mAs, 3–8 mm Voxel size: 1 × 1 × 5 mm3 dCCRT 87 (train) 46 (val.) HF In-house (Matlab 2015b) No Wang: 2019 [ 60 ] ESCC Train 2012–2016 Val. 2004–2014 CT 120 kV, 180-280 mA, 3 mm CCRT and RT alone 83 (train) 98 + 283 (val.)…”

“…The majority of imaging modalities mentioned in the retrieved studies were PET (20/37) [ 28 , 30 , 34 – 40 , 44 – 47 , 49 , 50 , 52 , 55 , 56 , 59 , 61 ], CT (16/37) [ 29 , 31 – 33 , 41 , 43 , 48 , 51 , 53 , 54 , 57 , 58 , 60 , 62 – 64 ], and one cone beam CT (CBCT) [ 42 ]. Although the search criteria included MRI, we did not locate any eligible study in our search.…”

“…The majority of patients included in 13 studies were treated specifically with radical CCRT [ 31 , 32 , 36 , 39 , 41 , 42 , 48 , 50 , 51 , 53 , 55 , 58 , 59 ]. In three studies, most patients were treated with CCRT, but the rest received a variety of different treatments depending on their situation [ 34 , 57 , 60 ]. There was one study that did not specify the intent of CCRT [ 37 ].…”

“…The number of patients reported in the included studies ranged from 20 [ 40 , 44 , 52 , 56 ] up to 464 [ 60 ]. Three studies utilized deep learning [ 46 , 53 , 64 ] and all other studies used only handcrafted features with Cox proportional hazards, logistic regression (LR), linear regression, support vector machine (SVM), and random forest (RF) models.…”