Page 130 - EJMO-9-1

Eurasian Journal of Medicine and
Oncology
Oncology care with AI chatbots and assistants

3.2.2. Data collection and integration module rather than replacing human expertise. Bejnordi et al.
13
The data collection and integration module collects patient emphasized the limitations of AI, highlighting the need for
data from electronic health records (EHRs) and integrates the human oversight in healthcare applications.
data into personalized profiles. Research by Gulshan et al. These components work together to enhance oncology
7
and Choi et al. highlight the significance of structured data care by leveraging AI-driven chatbots to address clinical
23
for tailoring responses and ensuring patient-specific support. and patient engagement needs while ensuring accuracy,
privacy, and compliance. The conceptualization of this block
3.2.3. NLP module diagram is also informed by the authors’ direct experience
The NLP Module processes patient inputs to understand in designing and analyzing AI-driven healthcare systems.
their intent and emotional state, enhancing communication Drawing on this expertise, the proposed framework
between patients and the chatbots. Esteva et al. showcased integrates evidence-based components with practical
6
NLP’s role in healthcare for interpreting patient needs, considerations for real-world implementation.
16
while Cirillo and Pippa emphasized its contribution to The proposed AI-driven chatbot system for oncology
improving patient-technology interaction. care can enhance patient interaction by integrating a series
3.2.4. Personalized response tailoring module of advanced features tailored to provide personalized
support. The interaction begins with patient inputs, such
The personalized response tailoring module uses ML as symptom descriptions or questions, which initiate the
algorithms to generate tailored responses. Liao et al. applied conversation. Using NLP, the system analyzes these inputs
8
AI methods to improve breast cancer diagnostics, which can to gauge the patient’s emotional state and needs. It then
extend to personalized chatbot interactions in oncology. pulls data from various sources, including EHRs, to create a
personalized treatment profile, ensuring that responses are
3.2.5. Proactive support module relevant to the patient’s specific oncology care needs. The
The proactive support module focuses on treatment system proactively supports patients by sending reminders
adherence by sending medication reminders and check- for treatment adherence and appointment scheduling.
ins. Chen et al. demonstrated the effectiveness of predictive Telemedicine capabilities further extend care by offering
9
models in enhancing adherence, guiding the development remote consultations and improving access for patients
of this module. in rural or underserved areas. ML models continuously
improve the system’s responses, incorporating the latest
3.2.6. Telemedicine integration module medical knowledge and adapting to individual patient
The telemedicine integration module extends care to remote needs. Data privacy and security are emphasized, ensuring
or underserved areas by facilitating virtual consultations. compliance with regulations, such as HIPAA to protect
Wang and Li highlighted AI’s role in bridging geographic sensitive patient information. In cases requiring more
10
disparities in oncology care. complex decisions, the system utilizes a human-in-
the-loop approach, escalating critical issues to human
3.2.7. Continuous learning and updates module oncologists. Finally, the system concludes by providing the
The continuous learning and updates module ensures patient with a personalized, empathetic response, ensuring
that the chatbots evolve by incorporating new medical that all interactions are both informative and supportive.
guidelines and patient interactions. Lee et al. emphasized 3.3. Sample input and output
17
the importance of adaptive AI systems that continuously
improve based on new data. The process of AI chatbot interaction in oncology involves
several key steps to ensure personalized and effective
3.2.8. Data security module support. Below is a structured explanation of the input and
The data security module ensures compliance with output flow, including the step-by-step interaction with the
regulations, such as the Health Insurance Portability and AI chatbot.
Accountability Act (HIPAA) and safeguarding patient i. Step 1: Patient input (Start patient interaction)
privacy. Mazurowski and Buda underscored the importance Patient query: “I’m feeling really nauseous after my
18
of robust data security in AI implementations in healthcare. chemotherapy. Do you have any suggestions to help
with this?”
3.2.9. Human-in-the-loop escalation module
ii. Step 2: Data Collection and Integration
The human-in-the-loop escalation module ensures complex Assistant action: The system retrieves relevant data
cases are escalated to human oncologists, complementing from the patient’s EHR, such as cancer type (breast

Volume 9 Issue 1 (2025) 122 doi: 10.36922/ejmo.6251