14.1.1. Background

Over the past few decades, advancements in early detection, targeted therapies, immunotherapy and multidisciplinary cancer care have led to a significant increase in cancer survival rates globally. While this is a remarkable achievement, survivorship comes with its own set of complex and long-term challenges (Mennella et al. 2024). Cancer survivors frequently contend with a spectrum of physical, neurological and psychological impairments that result from aggressive treatments such as chemotherapy, radiation therapy and surgical interventions. These may include cancer-related fatigue, chemotherapy-induced peripheral neuropathy, joint stiffness, muscular atrophy and restricted mobility. Such impairments can severely affect the survivor’s ability to perform daily activities and diminish the overall quality of life.

Physiotherapy is a cornerstone of post-treatment cancer care, aiming to restore functional independence, improve strength and endurance, and reduce the risk of secondary complications such as falls and muscle wasting (Bezzina 2023). However, traditional physiotherapy practices rely heavily on subjective observation and periodic assessments conducted during clinical visits. These methods are often insufficient for tracking subtle functional changes or adapting to the evolving needs of patients.

Additionally, logistical issues such as limited access to specialized care, patient nonadherence to home exercises and variability in clinical expertise present further barriers to effective rehabilitation.

The integration of AI, particularly through motion analysis technologies, offers a promising solution to these limitations (Phatak et al. 2021). AI-driven systems use computer vision, machine learning algorithms and sensor-based data to continuously and objectively monitor movement patterns. This allows for highly accurate analysis of joint kinematics, gait mechanics, posture and muscular coordination. The result is a more precise, personalized and adaptive rehabilitation approach that can respond in real time to a patient’s progress or regression.

14.1.2. Aim and scope

The aim of this research chapter is to explore the transformative role of AI-driven motion analysis in enhancing physiotherapy for cancer survivors (Al-Akayleh et al. 2024). Specifically, it investigates how AI technologies are being applied to overcome conventional physiotherapy challenges by providing continuous, quantitative and personalized monitoring of patient movement. The chapter systematically examines the core technologies underpinning AI-based motion analysis – including wearable sensors, computer vision platforms and deep learning algorithms – and their applications in both clinical and home-based rehabilitation settings.

14.2.1. Physical impairments in cancer survivors

Cancer treatments, while targeting the eradication of malignant cells, often have a profound impact on healthy tissues and organ systems (Kaur et al. 2023). These side effects can manifest as long-term or even permanent physical impairments that hinder survivors’ ability to engage in normal daily activities, maintain independence or return to work.

Some of the most commonly observed physical complications include:

• Lymphedema: a frequent consequence of lymph node dissection during breast, prostate or gynecologic cancer surgeries. This condition results in fluid accumulation, swelling and discomfort, typically in the limbs, which can significantly impair mobility and function.
  
• Peripheral neuropathy: caused by neurotoxic chemotherapeutic agents such as taxanes and platinum-based compounds (Was et al. 2022). Patients experience numbness, tingling, pain or weakness, primarily in the hands and feet, which affects balance, gait and fine motor skills.
  
• Reduced cardiovascular endurance: chemotherapy and radiation can weaken the heart muscle and vascular structures, contributing to fatigue and decreased exercise tolerance. Survivors often struggle to meet basic physical activity recommendations.
  
• Muscle atrophy and skeletal muscle loss (sarcopenia): prolonged bed rest, inactivity and catabolic effects of treatment lead to significant reductions in muscle mass and strength. This can increase the risk of falls, frailty and functional dependency.
  
• Joint stiffness and limited mobility: radiation fibrosis, scar formation after surgery or prolonged immobilization can lead to a decreased range of motion and joint contractures, affecting the ability to perform tasks such as reaching, bending or walking (Gomez et al. 2023).

Addressing these impairments requires a highly individualized rehabilitation approach that evolves with the patient’s recovery status. Traditional one-size-fits-all programs often fail to capture this dynamic need. Physiotherapists must monitor progress continuously, adjust therapy intensity and detect complications early; all of which are areas where AI-driven motion analysis systems can play a pivotal role.

By enabling accurate, real-time assessment of joint angles, muscle activation and gait mechanics, AI technologies can assist clinicians in fine-tuning interventions and detecting subtle deteriorations before they become clinically significant.

14.2.2. Current limitations in traditional physiotherapy

While physiotherapy is widely recognized as essential in cancer rehabilitation, conventional approaches are constrained by structural, methodological and logistical barriers (Navntoft et al. 2024).

Key limitations include:

• Reliance on manual goniometry and visual observation: in many settings, the assessment of joint range of motion, posture and balance is still performed manually. These methods are prone to human error, inter-observer variability and lack precision, especially when evaluating small changes in mobility or movement quality.
  
• Infrequent assessments due to limited clinician time: most patients attend physiotherapy sessions once or twice a week. Between sessions, their progress or decline may go unmonitored. This episodic care model fails to capture the continuous nature of rehabilitation, potentially delaying important therapy modifications.
  
• Patient noncompliance with home exercise programs: adherence to prescribed exercises is crucial for recovery but often compromised due to factors such as pain, lack of motivation, forgetfulness or misunderstanding of instructions (Room et al. 2021). Without effective monitoring, therapists cannot know whether patients are performing exercises correctly, or at all.
  
• Lack of data-driven personalization: rehabilitation plans are frequently based on standardized guidelines rather than real-time, patient-specific data. This can result in under-treatment or overexertion, both of which negatively impact recovery and patient satisfaction. Moreover, traditional physiotherapy lacks the predictive analytics to foresee complications or adjust for comorbidities dynamically.
  
• Limited access and resource constraints: in rural or underserved areas, access to specialized oncology physiotherapy services is scarce. Transportation issues, financial burdens and a shortage of trained professionals can further restrict continuity of care.

These constraints highlight a growing need for intelligent, scalable and patient-centered physiotherapy solutions. AI-driven motion analysis offers a compelling pathway to overcome many of these barriers (Reddy 2025). By enabling automated, accurate assessments and supporting personalized exercise plans, AI systems can empower both clinicians and patients to engage more effectively in the rehabilitation process, whether in clinics, homes or telehealth settings.

14.3.1. Motion capture systems

At the heart of AI-based physiotherapy systems lies the ability to accurately record and interpret human movement (Nambi et al. 2024). Several motion capture modalities are used to extract biomechanical data relevant to rehabilitation:

• Wearable sensors: devices such as inertial measurement units (IMUs), accelerometers, gyroscopes and magnetometers are affixed to key anatomical locations (e.g. knees, hips, spine) to measure linear acceleration, angular velocity and orientation. These systems offer mobility and can be used in both clinical and home settings.
  
• Computer vision systems: cameras – ranging from standard RGB to depth-sensing cameras such as Microsoft Kinect or LiDAR – track movements without the need for physical sensors. These systems can analyze posture, balance and dynamic motions in 3D, offering noninvasive, contactless monitoring.
  
• Markerless motion capture: recent advances in computer vision have enabled motion capture without traditional reflective markers. AI frameworks such as OpenPose, MediaPipe and PoseNet can estimate body keypoints from video footage, allowing for scalable and user-friendly analysis (Roggio et al. 2024). These technologies are particularly useful for tele-rehabilitation and remote assessments.

Together, these systems measure joint angles, movement velocity, stride length, limb symmetry and postural alignment. These are critical parameters for identifying impairments and tailoring therapy.

14.3.2. Machine learning and deep learning

Motion data collected from sensors and cameras is highly complex and multidimensional. To interpret this data meaningfully, AI systems use advanced machine learning and deep learning models. These models not only analyze current performance but also predict future trends and personalize interventions.

Key functionalities of AI models in this context include:

• Anomaly detection: AI can identify subtle abnormalities such as gait asymmetry, reduced joint flexion or poor balance that may be overlooked by the human eye (Harris et al. 2022). These insights are crucial for early intervention and progress tracking.
  
• Outcome prediction: machine learning algorithms can forecast patient outcomes based on historical and real-time data. For example, they can predict recovery times, risk of re-injury or adherence levels, helping clinicians make informed decisions.
  
• Adaptive therapy optimization: AI systems dynamically update rehabilitation protocols in response to patient progress. For instance, exercises may become more challenging as strength and coordination improve or be adjusted if regressions are detected.

Common models used include:

• Convolutional neural networks (CNNs): these are ideal for pose estimation and image-based motion tracking. CNNs process visual data to extract spatial patterns, such as joint locations or body alignment.
  
• Recurrent neural networks (RNNs) and long short-term memory (LSTM) networks
  Only gold members can continue reading. Log In or Register to continue

  Related posts:

  AI-driven Imaging in Oncology and Physiotherapy: A Game-changer in Diagnostics AI-enhanced Virtual Rehabilitation: The Future of Cancer Recovery Integrating AI with Physiotherapy to Improve Quality of Life for Cancer Patients Personalized AI-based Exercise Therapy for Oncology and Physiotherapy Patients AI-driven Innovations in Biomaterial-based Physiotherapy and Oncology: Personalizing Patient Care Through Intelligent Monitoring and Predictive Rehabilitation Enhancing Health Data Security and Intelligence with Blockchain and Artificial Intelligence
  

  Stay updated, free articles. Join our Telegram channel

  Join