Artificial intelligence (AI) promises to be the next revolutionary step in modern society. Yet, its role in all fields of industry and science need to be determined. One very promising field is represented by AI‐based decision‐making tools in clinical oncology leading to more comprehensive, personalized therapy approaches. In this review, the authors provide an overview on all relevant technical applications of AI in oncology, which are required to understand the future challenges and realistic perspectives for decision‐making tools. In recent years, various applications of AI in medicine have been developed focusing on the analysis of radiological and pathological images. AI applications encompass large amounts of complex data supporting clinical decision‐making and reducing errors by objectively quantifying all aspects of the data collected. In clinical oncology, almost all patients receive a treatment recommendation in a multidisciplinary cancer conference at the beginning and during their treatment periods. These highly complex decisions are based on a large amount of information (of the patients and of the various treatment options), which need to be analyzed and correctly classified in a short time. In this review, the authors describe the technical and medical requirements of AI to address these scientific challenges in a multidisciplinary manner. Major challenges in the use of AI in oncology and decision‐making tools are data security, data representation, and explainability of AI‐based outcome predictions, in particular for decision‐making processes in multidisciplinary cancer conferences. Finally, limitations and potential solutions are described and compared for current and future research attempts.

Purpose: Pre-operative assessment of surgical risk is essential for patient counselling in the elderly patient population. Our purpose was to compare validated geriatric assessment scores (GAS) in predicting postoperative morbidity and mortality in patients ≥ 80 years. Methods: Overall, eight preoperative GAS were assessed for each patient who received RC from 2016 to 2021. Postoperative morbidity was recorded according to the Clavien-Dindo classification (CDC) of surgical complications. Binary logistic regression analyses were used to determine prediction of 30-d morbidity and 90-d mortality in patients ≥ 80 years. Results: In total, 424 patients were analysed (77.4% male) with median age of 71 years (IQR: 68.82;70.69), of which 67 (15.8%) were ≥ 80 years. Patients age ≥ 80 years showed more 30-d CDC grade ≥ IIIb (41.07% vs. 27.74% compared to < 80 years, p < .001) and worse 90-d mortality (26.87% vs. 4.76%, p < .001). In patients ≥ 80 years, morbidity was predicted by simplified Frailty Index (sFI) ≥ 2 (OR: 2.06, 95% CI: 1.27–3.34, p = .004), Eastern Cooperative Oncology Group (ECOG) performance status ≥ 2 (OR: 2.78, 95% CI: 1.18–6.54, p = .019) and severe Adult Comorbidity Evaluation (ACE)-27 score (OR: 2.07, 95% CI: 1.13–3.79, p = .019), while 90-d mortality was predicted by CDC grade ≥ IIIb (OR: 22.91, 95% CI: 8.74–60.09, p < .001) and ECOG ≥ 2 (OR: 2.87, 95% CI: 1.05–7.86, p = .04). Conclusion: Even in a high-volume center of RC, 90-d mortality is significantly higher in patients age ≥ 80. Our results suggest in patient age ≥ 80, sFI ≥ 2, ECOG performance status ≥ 2 and severe ACE-27 score as clinical cut-off value to evaluate alternative bladder-sparing concepts.

Decisions on the best available treatment in clinical oncology are based on expert opinions in multidisciplinary cancer conferences (MCC). Artificial intelligence (AI) could increase evidence-based treatment by generating additional treatment recommendations (TR). We aimed to develop such an AI system for urothelial carcinoma (UC) and renal cell carcinoma (RCC). Comprehensive data of patients with histologically confirmed UC and RCC who received MCC recommendations in the years 2015 - 2022 were transformed into machine readable representations. Development of a two-step process to train a classifier to mimic TR was followed by identification of superordinate and detailed categories of TR. Machine learning (CatBoost, XGBoost, Random Forest) and deep learning (TabPFN, TabNet, SoftOrdering CNN, FCN) techniques were trained. Results were measured by F1-scores for accuracy weights. AI training was performed with 1617 (UC) and 880 (RCC) MCC recommendations (77 and 76 patient input parameters). The AI system generated fully automated TR with excellent F1-scores for UC (e.g. 'Surgery' 0.81, 'Anti-cancer drug' 0.83, 'Gemcitabine/Cisplatin' 0.88) and RCC (e.g. 'Anti-cancer drug' 0.92 'Nivolumab' 0.78, 'Pembrolizumab/Axitinib' 0.89). Explainability is provided by clinical features and their importance score. Finally, TR and explainability were visualized on a dashboard. This study demonstrates for the first time AI-generated, explainable TR in UC and RCC with excellent performance results as a potential support tool for high-quality, evidence-based TR in MCC. The comprehensive technical and clinical development sets global reference standards for future AI developments in MCC recommendations in clinical oncology. Next, prospective validation of the results is mandatory.

Purpose: Focal therapy (FT) for localized prostate cancer (PCa) is only recommended within the context of clinical trials by international guidelines. We aimed to investigate oncological follow-up and safety data of focal high-intensity focused ultrasound (HIFU) treatment. Methods: We conducted a single-center prospective study of 29 patients with PCa treated with (focal) HIFU between 2016 and 2021. Inclusion criteria were unilateral PCa detected by mpMRI-US-fusion prostate biopsy and maximum prostate specific antigen (PSA) of 15 ng/ml. Follow-up included mpMRI-US fusion-re-biopsies 12 and 24 months after HIFU. No re-treatment of HIFU was allowed. The primary endpoint was failure-free survival (FFS), defined as freedom from intervention due to cancer progression. Results: Median follow-up of all patients was 23 months, median age was 67 years and median preoperative PSA was 6.8 ng/ml. One year after HIFU treatment PCa was still detected in 13/ 29 patients histologically (44.8%). Two years after HIFU another 7/29 patients (24.1%) were diagnosed with PCa. Until now, PCa recurrence was detected in 11/29 patients (37.93%) which represents an FFS rate of 62%.One patient developed local metastatic disease 2 years after focal HIFU. Adverse events (AE) were low with 70% of patients remaining with sufficient erectile function for intercourse and 97% reporting full maintenance of urinary continence. Conclusion: HIFU treatment in carefully selected patients is feasible. However, HIFU was oncologically not as safe as expected because of progression rates of 37.93% and risk of progression towards metastatic disease. Thus, we stopped usage of HIFU in our department.

Background: In the treatment of advanced urothelial (aUC) and renal cell carcinoma (aRCC), biomarkers such as PD-1 and PD-L1 are not robust prognostic markers for immunotherapy (IO) response. Previously, a significant association between IO and a change in splenic volume (SV) was described for several tumour entities. To the best of our knowledge, this study presents the first correlation of SV to IO in aUC and aRCC. Methods: All patients with aUC (05/2017–10/2021) and aRCC (01/2012–05/2022) treated with IO at our academic centre were included. SV was measured at baseline, 3 and 9 months after initiation of IO using an in-house developed convolutional neural network-based spleen segmentation method. Uni- and multivariate Cox regression models for overall survival (OS) and progression-free survival (PFS) were used. Results: In total, 35 patients with aUC and 30 patients with aRCC were included in the analysis. Lower SV at the three-month follow-up was significantly associated with improved OS in the aRCC group. Conclusions: We describe a new, innovative artificial intelligence-based approach of a radiological surrogate marker for IO response in aUC and aRCC which presents a promising new predictive imaging marker. The data presented implicate improved OS with lower follow-up SV in patients with aRCC.