How Does AI Fit Into Hematologic Malignancy Management and Palliative Care?

As artificial intelligence (AI)–based tools become increasingly sophisticated and prevalent across clinical workflows, it is necessary to apply these technologies critically without replacing human judgment or decision-making skills, according to Ram Prakash Thirugnanasambandam, MBBS.

As part of a discussion centered around his publication The Role of Artificial Intelligence in Palliative Oncology: Zeroing in on Hematologic Malignancies, Thirugnanasambandam spoke with CancerNetwork^® about the utility that AI can provide in the context of palliative medicine and treatment of patients with different blood cancers. Thirugnanasambandam and colleagues published this review article in the November/December 2025 issue of the journal ONCOLOGY^®.

Thirugnanasambandam, a fellow of Internal Medicine, Hospice and Palliative Medicine, and Geriatric Medicine at the University of Miami, described AI’s ability to assist with identifying patients who might benefit from earlier palliative intervention and help manage symptoms related to cytopenias and infections. He ultimately emphasized that AI-based tools cannot substitute for human judgment or empathy in the clinic, pointing to a need for more refined datasets to elevate patient care.

CancerNetwork: What was the rationale for conducting a systematic review to assess the use of AI in palliative care and hematologic malignancies?

Thirugnanasambandam: [Our] idea came from a practical standpoint. I was a palliative care fellow at the University of Nebraska last year. During [this] year, I continuously heard the term AI being brought up in lectures, conferences, and informal discussions with the mentors. That prompted me to take a step back and ask a bigger question: what do we know about AI so far? Particularly, in the context of palliative care, how do we utilize it in terms of patient-centered goals, whether it's symptom management, communication, or decision-making support? [Additionally], I had an interest in oncology.

We decided to focus on hematologic malignancies particularly because these patients tend to have unpredictable disease trajectories as well as complex symptom burden. The main aim of this review article was to bring the palliative care and hematologic malignancy strands together. [Our review aimed to] help clinicians identify what is currently known, clarify what is clinically relevant, and help highlight where further caution and validation is required.

What did data show regarding the role that AI could play in palliative care? How can AI-based tools help enhance symptom management and support decision-making in these patient populations?

What the data showed us is that in palliative care, AI is best used as a supportive tool. It cannot replace clinical judgment or decision-making, but it can help flag patients who might benefit from earlier palliative care interventions. In terms of symptom management, the ability of AI to do predictive analysis based on large amounts of data is key to palliative care because by looking at large amounts of data, it can now identify patients who may have symptoms such as pain, dyspnea, anxiety, or psychosocial distress, allowing us to take a more proactive approach to patient care. In terms of decision-making, I would say AI is currently acting only as a prompt because it can [cue] clinicians to have serious illness conversations or involve [the] palliative care team earlier.

How might AI-based tools show utility in hematologic malignancies, especially as they relate to areas like diagnostics, symptom management, and prognostication and risk stratification?

Hematologic malignancy is an area where AI has shown relatively mature applications, mostly because these diseases have a large amount of data that can be analyzed with the use of AI-based tools. Whether it's in use of pathology slides, blood smears, longitudinal clinical data, or imaging studies, these diseases tend to produce a lot of data that can be utilized by AI. In terms of a diagnostic standpoint, studies have shown that by utilizing the pathology smears, the blood smears, and PET/CT imaging, these AI-based tools can accurately predict disease subtypes as well as the disease burden in hematologic malignancies, whether it's leukemia, lymphoma, or [multiple] myeloma.

From a symptom standpoint, it's able to [conduct] predictive analysis, particularly in hematologic malignancies. Patients can have symptom management, which is difficult because they have fluctuating and more severe symptoms whether they are related to cytopenias, transfusion-related needs, infections, or even treatment-related toxicities. [AI is] able to predict when a patient may have anemia or may experience infection, allowing clinicians to have a more proactive approach in managing these symptoms. From a prognostication and risk stratification [perspective], we've used limited clinical variables or genetic variables. However, by having a large amount of data from healthcare records, it's able to provide a more personalized approach.

What are some potential ethical considerations and limitations associated with AI tools in these patient populations? What should clinicians and researchers keep in mind to use AI as responsibility as possible?

One of the main themes that emerged while we were doing this review article is that, yes, AI can be utilized well in palliative care and [hematologic] malignancies, but we also [must] think about the ethical considerations while utilizing it. From a patient perspective, that can be related to data privacy consent from patients as well as bias. For example, one of the studies we looked at in the review article showed AI-based tool healthcare utilization needs in comparison to true needs for a particular patient population, and that may not be reflective of all patients across the country. That can reflect systemic inequities rather than true need of care in these patient populations.

Another standpoint is that while AI can be utilized as a decision-making support, it cannot replace nuanced clinical judgment and empathy in terms of palliative care. It can [conduct] predictive analysis, but it cannot predict the values, goals, and emotional distress that a patient goes through while they go through these difficult diseases. These are ethical considerations that need to take into the bigger picture and have transparency and continuous clinical oversight from a human perspective. We also need to train these AI models on different data—data from diverse patient populations—so that there's no bias when it comes to the utilization of AI.

What are the next steps for further improving outcomes in palliative care and among patients with hematologic cancers?

Looking ahead, while we continue to [observe] at data from a retrospective and observational standpoint, we also have to design clinical trials where we can prospectively look at how AI is improving the quality of care that we provide to patients. In terms of hematologic malignancies, we need to continuously refine risk stratification and prognostication. In terms of data bias, like I mentioned before, we need to continuously train these models so that we can avoid bias and provide equitable care to all patient populations across the country. Having clinical oversight is extremely important when utilizing AI so that we can train clinicians to use it more thoughtfully as well as critically. [Additionally], when we utilize AI, we need to make sure it doesn't increase the burden placed on clinicians. By doing that, we can allow clinicians to spend more time with patients as opposed to doing administrative tasks.

Looking ahead, what other roles or purposes do you see AI and machine learning fulfilling in palliative medicine, hematologic malignancies, and other oncologic populations?

The evolving role [of AI] within palliative care is going to be mostly in terms of supportive management. Like I mentioned before, it cannot replace clinical judgment or decision-making for clinicians, but it can help flag patients. That's where I see its utility the most. In terms of hematologic malignancies, the biggest evolving factor is its use in survivorship care because these patients may have continuous functional needs, supportive care needs, or palliative care needs. It can also track late symptoms that can occur in patients.

What do you hope others take away from your publication?

I want readers to take away a sense of balance. We’ve done the article to help clinicians be more comfortable in engaging with AI. We need to apply it critically, not as replacing judgment or decision-making skills, but more as an adjunct. If we’re able to spark meaningful conversation, we’ve done our job rather than just hyping up AI.