New AI Tool Revolutionizes Diagnosis of Muscle Wasting in Head and Neck Cancer Patients

An innovative artificial intelligence (AI) tool developed by researchers at the Dana-Farber Cancer Institute is set to revolutionize the diagnosis of muscle wasting, known as sarcopenia, in patients with head and neck cancer. The tool, designed to provide a fast and accurate assessment, offers a significant improvement over the current time-consuming and laborious manual diagnosis process conducted by human experts.

Sarcopenia is a strong indicator of a patient’s deteriorating health, with significant implications for treatment and supportive care. “A real-time tool that tells us when a patient is losing muscle mass would trigger us to intervene and do something supportive to help,” explains Benjamin Kann, MD, lead author and radiation oncologist.

While head and neck cancer treatments can be effective, they often come with severe side effects that impact a patient’s ability to eat and drink, leading to poor nutrition and muscle wasting. Sarcopenia is associated with a higher likelihood of requiring feeding tubes, experiencing a lower quality of life, and worse overall outcomes, including premature death.

Traditionally, muscle mass assessments have been conducted through computed tomography (CT) scans of the abdomen or neck. However, the process of analyzing these scans to identify sarcopenia is time-consuming and requires highly skilled experts, making it impractical for routine use.

To address this limitation, Kann and his colleagues turned to deep learning, a branch of AI, to develop a model that can automatically diagnose sarcopenia by accurately distinguishing muscle from other tissues in CT scans of the neck. The team trained the model using clinical records and CT scans from 420 head and neck cancer patients, with an expert calculating a skeletal muscle index (SMI) score for each patient based on the scans.

The AI model successfully delineated the muscle in the neck and provided transparent, verifiable results. In a validation study, the model demonstrated a 96.2% accuracy rate in assessing muscle mass, as confirmed by an expert panel review. Moreover, the AI model can complete the assessment in just 0.15 seconds, significantly reducing the time required for diagnosis.

Compared to the widely used body mass index (BMI), SMI was found to be a superior predictor of poor outcomes, making it a more valuable clinical tool for monitoring patient health during treatment. By regularly assessing sarcopenia, physicians can detect muscle mass decline early and implement interventions such as nutritional consultations, supportive medications, or physical therapy to prevent further deterioration.

In addition to monitoring patient decline, the AI tool has the potential to guide treatment decisions from the outset. Patients diagnosed with cancer who already exhibit sarcopenia may benefit from a less aggressive treatment approach compared to those with higher muscle mass.

Moving forward, Kann and his team plan to further explore the tool’s capabilities by applying it in a clinical trial setting. By continuously monitoring muscle mass changes throughout treatment, they hope to gain insights that will inform treatment decisions and interventions, ultimately improving patient outcomes.

FAQ:

Q: What is sarcopenia?
A: Sarcopenia is the progressive loss of muscle mass, strength, and function associated with aging or certain medical conditions.

Q: Why is sarcopenia important in head and neck cancer patients?
A: Sarcopenia in head and neck cancer patients can lead to poorer outcomes, including the need for feeding tubes, lower quality of life, and increased mortality rates.

Q: How does the AI tool work?
A: The AI tool uses deep learning algorithms to analyze CT scans of the neck and automatically detect and assess muscle mass, providing a fast and accurate diagnosis of sarcopenia.

Q: How does the AI tool compare to conventional methods?
A: The AI tool significantly reduces the time required for diagnosis, providing results in just 0.15 seconds compared to the manual process, which takes up to 10 minutes. It also eliminates the need for highly trained experts to manually analyze the scans, making it a practical and efficient diagnostic tool.

Q: Can the AI tool guide treatment decisions?
A: Yes, the AI tool can assist in treatment decision-making by identifying sarcopenia early on. This information allows physicians to tailor treatments and interventions according to a patient’s muscle mass, potentially improving treatment outcomes.

Source: Dana-Farber Cancer Institute