Digital Guardian Angel Supported by an Artificial Intelligence System to Improve Quality of Life, Well-being, and Health Outcomes of Patients With Cancer (ONCORELIEF): Protocol for a Single Arm Prospective Multicenter Pilot Study.

BACKGROUND: According to Europe's Beating Cancer Plan, the number of cancer survivors is growing every year and is now estimated at over 12 million in Europe. A main objective of the European Commission is to ensure that cancer survivors can enjoy a high quality of life, underlining the role of digital technology and eHealth apps and tools to achieve this. OBJECTIVE: The main objective of this study is the development of a user-centered artificial intelligence system to facilitate the input and integration of patient-related biopsychosocial data to improve posttreatment quality of life, well-being, and health outcomes and examine the feasibility of this digitally assisted workflow in a real-life setting in patients with colorectal cancer and acute myeloid leukemia. METHODS: A total of 60 patients with colorectal cancer and 30 patients with acute myeloid leukemia will be recruited from 2 clinical centers: Universitätsmedizin der Johannes Gutenberg-Universität Mainz (Mainz, Germany) and IRCCS Istituto Romagnolo per lo Studio dei Tumori "Dino Amadori" (IRST, Italy). Psychosocial data (eg, emotional distress, fatigue, quality of life, subjective well-being, sleep problems, and appetite loss) will be collected by questionnaires via a smartphone app, and physiological data (eg, heart rate, skin temperature, and movement through step count) will be collected by a customizable smart wrist-worn sensor device. Each patient will be assessed every 2 weeks over their 3-month participation in the ONCORELIEF study. Inclusion criteria include patients with the diagnosis of acute myeloid leukemia or colorectal cancer, adult patients aged 18 years and older, life expectancy greater than 12 months, Eastern Cooperative Oncology Group performance status ≤2, and patients who have a smartphone and agree to use it for the purpose of the study. Exclusion criteria include patients with a reduced cognitive function (such as dementia) or technological illiteracy and other known active malignant neoplastic diseases (patients with a medical history of treated neoplastic disease are included). RESULTS: The pilot study started on September 1, 2022. As of January 2023, we enrolled 33 patients with colorectal cancer and 7 patients with acute myeloid leukemia. As of January 2023, we have not yet started the data analysis. We expect to get all data in June 2023 and expect the results to be published in the second semester of 2023. CONCLUSIONS: Web-based and mobile apps use methods from mathematical decision support and artificial intelligence through a closed-loop workflow that connects health professionals and patients. The ONCORELIEF system has the potential of continuously identifying, collecting, and processing data from diverse patient dimensions to offer health care recommendations, support patients with cancer to address their unmet needs, and optimize survivorship care. TRIAL REGISTRATION: German Clinical Trials Register (DRKS) 00027808; https://drks.de/search/en/trial/DRKS00027808. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/45475.

A new mathematical approach for handling DVH criteria in IMRT planning.

The appropriate handling of planning criteria on the cumulative dose-volume histogram (DVH) is a highly problematic issue in intensity-modulated radiation therapy (IMRT) plan optimization. The nonconvexity of DVH criteria and globality of the resulting optimization problems complicate the design of suitable optimization methods, which feature numerical efficiency, reliable convergence and optimality of the results. This work examines the mathematical structure of DVH criteria and proves the valuable properties of isotonicity/antitonicity, connectedness, invexity and sufficiency of the KKT condition. These properties facilitate the use of efficient and goal-oriented optimization methods. An exemplary algorithmic realization with feasible direction methods gives rise to a functional framework for interactive IMRT planning on DVH criteria. Numerical examples on real world planning cases prove its practical capability.

Analyzing the quality robustness of chemotherapy plans with respect to model uncertainties.

Mathematical models of chemotherapy planning problems contain various biomedical parameters, whose values are difficult to quantify and thus subject to some uncertainty. This uncertainty propagates into the therapy plans computed on these models, which poses the question of robustness to the expected therapy quality. This work introduces a combined approach for analyzing the quality robustness of plans in terms of dosing levels with respect to model uncertainties in chemotherapy planning. It uses concepts from multi-criteria decision making for studying parameters related to the balancing between the different therapy goals, and concepts from sensitivity analysis for the examination of parameters describing the underlying biomedical processes and their interplay. This approach allows for a profound assessment of a therapy plan, how stable its quality is with respect to parametric changes in the used mathematical model.

Breast cancer therapy planning - a novel support concept for a sequential decision making problem.

Breast cancer is the most common carcinosis with the largest number of mortalities in women. Its therapy comprises a wide spectrum of different treatment modalities a breast oncologist decides about for the individual patient case. These decisions happen according to medical guide lines, current scientific publications and experiences acquired in former cases. Clinical decision making therefore involves the time-consuming search for possible therapy options and their thorough testing for applicability to the current patient case.This research work addresses breast cancer therapy planning as a multi-criteria sequential decision making problem. The approach is based on a data model for patient cases with therapy descriptions and a mathematical notion for therapeutic relevance of medical information. This formulation allows for a novel decision support concept, which targets at eliminating observed weaknesses in clinical routine of breast cancer therapy planning.

The integration of DVH-based planning aspects into a convex intensity modulated radiation therapy optimization framework.

The formulation of intensity modulated radiation therapy (IMRT) planning aspects frequently uses the dose-volume histogram (DVH), whereas plan computations often happen in the more desirable convex IMRT optimization framework. Inspired by a recent publication of Zinchenko et al (2008 Phys. Med. Biol. 53 3231-50), this work addresses the integration of DVH-based planning aspects into this framework from a general point of view. It first provides the basic mathematical requirements on the evaluation functions in order to support such an incorporation. Then it introduces the condition number as a description for how precisely DVH-based planning aspects can be reformulated in terms of evaluation functions. Exemplary numerical studies for the generalized equivalent uniform dose and a physical constraint function show the influence of function parameter values and DVH approximation on the condition number. The work concludes by formulating the aspects that should be taken into account for an appropriate integration of DVH-based planning aspects.

On the role of modeling parameters in IMRT plan optimization.

The formulation of optimization problems in intensity-modulated radiotherapy (IMRT) planning comprises the choice of various values such as function-specific parameters or constraint bounds. In current inverse planning programs that yield a single treatment plan for each optimization, it is often unclear how strongly these modeling parameters affect the resulting plan. This work investigates the mathematical concepts of elasticity and sensitivity to deal with this problem. An artificial planning case with a horse-shoe formed target with different opening angles surrounding a circular risk structure is studied. As evaluation functions the generalized equivalent uniform dose (EUD) and the average underdosage below and average overdosage beyond certain dose thresholds are used. A single IMRT plan is calculated for an exemplary parameter configuration. The elasticity and sensitivity of each parameter are then calculated without re-optimization, and the results are numerically verified. The results show the following. (1) elasticity can quantify the influence of a modeling parameter on the optimization result in terms of how strongly the objective function value varies under modifications of the parameter value. It also can describe how strongly the geometry of the involved planning structures affects the optimization result. (2) Based on the current parameter settings and corresponding treatment plan, sensitivity analysis can predict the optimization result for modified parameter values without re-optimization, and it can estimate the value intervals in which such predictions are valid. In conclusion, elasticity and sensitivity can provide helpful tools in inverse IMRT planning to identify the most critical parameters of an individual planning problem and to modify their values in an appropriate way.

A new concept for interactive radiotherapy planning with multicriteria optimization: first clinical evaluation.

BACKGROUND AND PURPOSE: Currently, inverse planning for intensity-modulated radiotherapy (IMRT) can be a time-consuming trial and error process. This is because many planning objectives are inherently contradictory and cannot reach their individual optimum all at the same time. Therefore in clinical practice the potential of IMRT cannot be fully exploited for all patients. Multicriteria (multiobjective) optimization combined with interactive plan navigation is a promising approach to overcome these problems. PATIENTS AND METHODS: We developed a new inverse planning system called "Multicriteria Interactive Radiotherapy Assistant (MIRA)". The optimization result is a database of patient specific, Pareto-optimal plan proposals. The database is explored with an intuitive user interface that utilizes both a new interactive element for plan navigation and familiar dose visualizations in form of DVH and isodose projections. Two clinical test cases, one paraspinal meningioma case and one prostate case, were optimized using MIRA and compared with the clinically approved planning program KonRad. RESULTS: Generating the databases required no user interaction and took approx. 2-3h per case. The interactive exploration required only a few minutes until the best plan was identified, resulting in a significant reduction of human planning time. The achievable plan quality was comparable to KonRad with the additional benefit of having plan alternatives at hand to perform a sensitivity analysis or to decide for a different clinical compromise. CONCLUSIONS: The MIRA system provides a complete database and interactive exploration of the solution space in real time. Hence, it is ideally suited for the inherently multicriterial problem of inverse IMRT treatment planning.

IMRT planning on adaptive volume structures--a decisive reduction in computational complexity.

The objective of radiotherapy planning is to find a compromise between the contradictive goals of delivering a sufficiently high dose to the target volume while widely sparing critical structures. The search for such a compromise requires the computation of several plans, which mathematically means solving several optimization problems. In the case of intensity modulated radiotherapy (IMRT) these problems are large-scale, hence the accumulated computational expense is very high. The adaptive clustering method presented in this paper overcomes this difficulty. The main idea is to use a preprocessed hierarchy of aggregated dose-volume information as a basis for individually adapted approximations of the original optimization problems. This leads to a decisively reduced computational expense: numerical experiments on several sets of real clinical data typically show computation times decreased by a factor of about 10. In contrast to earlier work in this field, this reduction in computational complexity will not lead to a loss in accuracy: the adaptive clustering method produces the optimum of the original optimization problem.