Adaptive Treatment of Metastatic Prostate Cancer Using Generative Artificial Intelligence

Despite the expanding therapeutic options available to cancer patients, therapeutic resistance, disease recurrence, and metastasis persist as hallmark challenges in the treatment of cancer. The rise to prominence of generative artificial intelligence (GenAI) in many realms of human activities is compelling the consideration of its capabilities as a potential lever to advance the development of effective cancer treatments. This article presents a hypothetical case study on the application of generative pre-trained transformers (GPTs) to the treatment of metastatic prostate cancer (mPC). The case explores the design of GPT-supported adaptive intermittent therapy for mPC. Testosterone and prostate-specific antigen (PSA) are assumed to be repeatedly monitored while treatment may involve a combination of androgen deprivation therapy (ADT), androgen receptor-signalling inhibitors (ARSI), chemotherapy, and radiotherapy. The analysis covers various questions relevant to the configuration, training, and inferencing of GPTs for the case of mPC treatment with a particular attention to risk mitigation regarding the hallucination problem and its implications to clinical integration of GenAI technologies. The case study provides elements of an actionable pathway to the realization of GenAI-assisted adaptive treatment of metastatic prostate cancer. As such, the study is expected to help facilitate the design of clinical trials of GenAI-supported cancer treatments.

How higher goals are constructed and collapse under stress: A hierarchical Bayesian control systems perspective

In this paper, we show that organisms can be modeled as hierarchical Bayesian control systems with small world and information bottleneck (bow-tie) network structure. Such systems combine hierarchical perception with hierarchical goal setting and hierarchical action control. We argue that hierarchical Bayesian control systems produce deep hierarchies of goal states, from which it follows that organisms must have some form of 'highest goals'. For all organisms, these involve internal (self) models, external (social) models and overarching (normative) models. We show that goal hierarchies tend to decompose in a top-down manner under severe and prolonged levels of stress. This produces behavior that favors short-term and self-referential goals over long term, social and/or normative goals. The collapse of goal hierarchies is universally accompanied by an increase in entropy (disorder) in control systems that can serve as an early warning sign for tipping points (disease or death of the organism). In humans, learning goal hierarchies corresponds to personality development (maturation). The failure of goal hierarchies to mature properly corresponds to personality deficits. A top-down collapse of such hierarchies under stress is identified as a common factor in all forms of episodic mental disorders (psychopathology). The paper concludes by discussing ways of testing these hypotheses empirically.

LSD1-, EDS1- and PAD4-dependent conditional correlation among salicylic acid, hydrogen peroxide, water use efficiency and seed yield in Arabidopsis thaliana

In Arabidopsis thaliana, LESION SIMULATING DISEASE 1 (LSD1), ENHANCED DISEASE SUSCEPTIBILITY 1 (EDS1) and PHYTOALEXIN DEFICIENT 4 (PAD4) proteins are regulators of cell death (CD) in response to abiotic and biotic stresses. Hormones, such as salicylic acid (SA), and reactive oxygen species, such as hydrogen peroxide (H₂ O₂ ), are key signaling molecules involved in plant CD. The proposed mathematical models presented in this study suggest that LSD1, EDS1 and PAD4 together with SA and H₂ O₂ are involved in the control of plant water use efficiency (WUE), vegetative growth and generative development. The analysis of Arabidopsis wild-type and single mutants lsd1, eds1, and pad4, as well as double mutants eds1/lsd1 and pad4/lsd1, demonstrated the strong conditional correlation between SA/H₂ O₂ and WUE that is dependent on LSD1, EDS1 and PAD4 proteins. Moreover, we found a strong correlation between the SA/H₂ O₂ homeostasis of 4-week-old Arabidopsis leaves and a total seed yield of 9-week-old plants. Altogether, our results prove that SA and H₂ O₂ are conditionally regulated by LSD1/EDS/PAD4 to govern WUE, biomass accumulation and seed yield. Conditional correlation and the proposed models presented in this study can be used as the starting points in the creation of a plant breeding algorithm that would allow to estimate the seed yield at the initial stage of plant growth, based on WUE, SA and H₂ O₂ content.

The Adaptive Complexity of Cancer

Cancer treatment options are expanding to the benefit of significant segments of patients. However, their therapeutic power is not equally realized for all cancer patients due to drug toxicity and disease resistance. Overcoming these therapeutic challenges would require a better understanding of the adaptive survival mechanisms of cancer. In this respect, an integrated view of the disease as a complex adaptive system is proposed as a framework to explain the dynamic coupling between the various drivers underlying tumor growth and cancer resistance to therapy. In light of this system view of cancer, the immune system is in principal the most appropriate and naturally available therapeutic instrument that can thwart the adaptive survival mechanisms of cancer. In this respect, new cancer therapies should aim at restoring immunosurveillance by priming the induction of an effective immune response through a judicious targeting of immunosuppression, inflammation, and the tumor nutritional lifeline extended by the tumor microenvironment.

Perspective on the dynamics of cancer

BACKGROUND

The genetic diversity of cancer and the dynamic interactions between heterogeneous tumor cells, the stroma and immune cells present daunting challenges to the development of effective cancer therapies. Although cancer biology is more understood than ever, this has not translated into therapies that overcome drug resistance, cancer recurrence and metastasis. The future development of effective therapies will require more understanding of the dynamics of homeostatic dysregulation that drives cancer growth and progression.

RESULTS

Cancer dynamics are explored using a model involving genes mediating the regulatory interactions between the signaling and metabolic pathways. The exploration is informed by a proposed genetic dysregulation measure of cellular processes. The analysis of the interaction dynamics between cancer cells, cancer associated fibroblasts, and tumor associate macrophages suggests that the mutual dependence of these cells promotes cancer growth and proliferation. In particular, MTOR and AMPK are hypothesized to be concurrently activated in cancer cells by amino acids recycled from the stroma. This leads to a proliferative growth supported by an upregulated glycolysis and a tricarboxylic acid cycle driven by glutamine sourced from the stroma. In other words, while genetic aberrations ignite carcinogenesis and lead to the dysregulation of key cellular processes, it is postulated that the dysregulation of metabolism locks cancer cells in a state of mutual dependence with the tumor microenvironment and deepens the tumor's inflammation and immunosuppressive state which perpetuates as a result the growth and proliferation dynamics of cancer.

CONCLUSIONS

Cancer therapies should aim for a progressive disruption of the dynamics of interactions between cancer cells and the tumor microenvironment by targeting metabolic dysregulation and inflammation to partially restore tissue homeostasis and turn on the immune cancer kill switch. One potentially effective cancer therapeutic strategy is to induce the reduction of lactate and steer the tumor microenvironment to a state of reduced inflammation so as to enable an effective intervention of the immune system. The translation of this therapeutic approach into treatment regimens would however require more understanding of the adaptive complexity of cancer resulting from the interactions of cancer cells with the tumor microenvironment and the immune system.

State machine modeling of MAPK signaling pathways

Mitogen activated protein kinase (MAPK) signaling pathways are frequently deregulated in human cancers with potential involvement in most if not all cellular processes leading to tumorigenesis. Mathematical/computational models of MAPK signaling are indispensable to the study of pathway deregulation dynamics and their nonlinear effects on cell fate and carcinogenesis. A finite state machine model of MAPK cellular signaling is explored as an alternative to differential equations-based models of kinetics. The proposed approach is applied to the Ras-Extracellular signal-regulated kinase (Ras-ERK) pathway which includes the frequently mutated Ras and RAF proteins in many types of carcinomas.