Alain Hendlisz

PURPOSE: Managing chemorefractory metastatic colorectal cancer (mCRC) requires a meticulous equilibrium between the efficacy and toxicity of interventions, a task compounded by the constrained life expectancy of the patient. While existing prognostic tools, such as the Colon Life nomogram, primarily focus on general patient conditions or a single diagnostic modality, they do not fully integrate the potential predictive value of multimodal data. This study aims to develop and validate an Imaging Score, integrating clinical and imaging features derived from whole-body ¹⁸ F-fluorodeoxyglucose ( ¹⁸ F-FDG) positron emission tomography-computed tomography (PET-CT), predicting death probability within 12 weeks from treatment initiation for refractory disease. MATERIALS AND METHODS: The development cohort comprises 254 patients from three clinical trials. Nine clinical variables and six imaging variables were assessed. After optimal subset selection through recursive Feature Elimination with cross-validation, a support vector classifier-trained machine learning model generated the Imaging Score. Validation was performed on a real-life patient cohort (n = 74). Model performance was assessed on discrimination (Harrell C-index) and calibration. RESULTS: Final prognostic features included whole-body metabolically active tumor volume, Eastern Cooperative Oncology Group performance status, visceral fat density, number of metastatic sites, body mass index, maximum standardized distance, and months since diagnosis. The Imaging Score demonstrated robust discriminative ability in both the development (C-index, 0.797) and validation (C-index, 0.714) sets, outperforming the Colon Life nomogram that tended to overestimate 12-week mortality. CONCLUSION: The Imaging Score, integrating ¹⁸ F-FDG PET-CT imaging with clinical parameters, is an effective prognostic tool for patients with chemorefractory mCRC. This combination of imaging biomarkers with clinical factors improves discrimination, enhancing its potential for clinical decision making, patient stratification for chemorefractory treatments, and trial eligibility.

Immune checkpoint inhibitors and immune-related biomarkers are increasingly investigated in rectal cancer (RC). We retrospectively analysed PD-L1 expression in diagnostic biopsy and resection samples from RC patients treated at our centre between 2000 and 2020. PD-L1 immunostaining (22C3 clone) was evaluated according to tumour proportion (TPS), immune cell (ICS), and the combined positive score (CPS). Eighty-three patients were included. At diagnosis, PD-L1 expression ≥1%/≥5% was observed in 15.4%/0%, 80.7%/37.4%, and 69.2%/25.6% of patients based on TPS, ICS, and CPS, respectively. At surgery, the respective figures were 4.6%/1.5%, 60.2%/32.5%, and 50.7%/26.2%. Using the 1% cut-off and regardless of the scoring system, PD-L1 was less expressed in surgery than biopsy samples (p ≤ 0.04). In paired specimens, PD-L1-ICS reduction was especially observed following neoadjuvant long-course (chemo)radiotherapy (p = 0.03). PD-L1-ICS of ≥5% in surgical samples (HR: 0.17; p = 0.02), and a biopsy-to-surgery increase in PD-L1-ICS (HR: 0.19; p = 0.04) was predictive for longer disease-free survival, while the PD-L1-ICS of either ≥1% (HR 0.28; p = 0.04) or ≥5% (HR 0.19; p = 0.03) in surgical samples and the biopsy-to-surgery increase in PD-L1-ICS (HR: 0.20; p = 0.04) were associated with better overall survival. Our study suggests that PD-L1 expression in RC is largely reflective of immune cell infiltration, and its presence/increase in surgical samples predicts better outcomes.

Purpose. This multicenter retrospective study aims to identify reliable clinical and radiomic features to build machine learning models that predict progression-free survival (PFS) and overall survival (OS) in pancreatic ductal adenocarcinoma (PDAC) patients. Methods. Between 2010 and 2020 pre-treatment contrast-enhanced CT scans of 287 pathology-confirmed PDAC patients from two sites of the Hopital Universitaire de Bruxelles (HUB) and from 47 hospitals within the HUB network were retrospectively analysed. Demographic, clinical, and survival data were also collected. Gross tumour volume (GTV) and non-tumoral pancreas (RPV) were semi-manually segmented and radiomics features were extracted. Patients from two HUB sites comprised the training dataset, while those from the remaining 47 hospitals of the HUB network constituted the testing dataset. A three-step method was used for feature selection. Based on the GradientBoostingSurvivalAnalysis classifier, different machine learning models were trained and tested to predict OS and PFS. Model performances were assessed using the C-index and Kaplan–Meier curves. SHAP analysis was applied to allow for post hoc interpretability. Results. A total of 107 radiomics features were extracted from each of the GTV and RPV. Fourteen subgroups of features were selected: clinical, GTV, RPV, clinical & GTV, clinical & GTV & RPV, GTV-volume and RPV-volume both for OS and PFS. Subsequently, 14 Gradient Boosting Survival Analysis models were trained and tested. In the testing dataset, the clinical & GTV model demonstrated the highest performance for OS (C-index: 0.72) among all other models, while for PFS, the clinical model exhibited a superior performance (C-index: 0.70). Conclusions. An integrated approach, combining clinical and radiomics features, excels in predicting OS, whereas clinical features demonstrate strong performance in PFS prediction.

Circulating cytokines could be optimal biomarkers for prognostication and management decisions in colorectal cancer (CRC). Chemorefractory CRC patients with available plasma samples were included in this study. In the discovery cohort (n = 85), 182 circulating cytokines were tested with a semi-quantitative multiplex assay, and prognostic cytokines were analyzed in the validation cohort (n = 111) by ELISA. Overall survival (OS) was the primary outcome measure, with the false discovery rate (FDR) method (significance level of <0.01) being used to correct for multiple comparisons. Four cytokines were associated with OS in the discovery cohort: insulin-like growth factor-binding protein 1 (IGFBP-1) (HR 2.1 [95%CI: 1.58–2.79], FDR < 0.001), insulin-like growth factor-binding protein 2 (IGFBP-2) (HR 1.65 [95%CI: 1.28–2.13], FDR = 0.006), serum amyloid A (SAA) (HR 1.84 [95%CI: 1.39–2.43], FDR < 0.001), and angiotensin II (HR 1.65 [95%CI: 1.29–2.1], FDR = 0.006). Of these, IGFBP-1 (HR 2.70 [95%CI: 1.56–4.76], FDR = 0.007) and IGFBP-2 (HR 3.33 [95%CI: 1.64–6.67], FDR = 0.008) were confirmed to be independently associated with OS in the validation cohort. Patients with high concentrations of IGFBP-1 and/or IGFBP-2 had a median OS of 3.0 months as compared with 6.9 months for those with low concentrations of both cytokines (HR 2.44 [95%CI: 1.52–4.0], FDR = 0.002) Validation of circulating IGFBP-1 and IGFBP-2 as independent prognostic biomarkers for chemorefractory CRC in larger, independent series is warranted.

Potential intrinsic resistance mechanisms to regorafenib were explored after short exposure (3 days) on five CRC cell lines (HCT-116, SW1116, LS-1034, SW480, Caco-2). The observation of senescence-like features led to the investigation of a drug-initiated phenotype switch. Following long-term exposure (12 months) of HCT-116 and SW480 cell lines to regorafenib, we developed resistant models to explore acquired resistance. SW480 cells demonstrated senescent-like properties, including a cell arrest in the late G2/prophase cell cycle stage and a statistically significant decrease in the expression of G1 Cyclin-Dependent Kinase inhibitors and key cell cycle regulators. A specific senescence-associated secretome was also observed. In contrast, HCT-116 treated cells presented early senescent features and developed acquired resistance triggering EMT and a more aggressive phenotype over time. The gained migration and invasion ability by long-exposed cells was associated with the increased expression level of key cellular and extracellular EMT-related factors. The PI3K/AKT pathway was a significant player in the acquired resistance of HCT-116 cells, possibly related to a PI3KCA mutation in this cell line. Our findings provide new insights into the phenotypic plasticity of CRC cells able, under treatment pressure, to acquire a stable TIS or to use an early senescence state to undergo EMT.