Page 80 - AIH-1-4

Artificial Intelligence in Health Robotics and Vivaldi AI for ALS assessment

1. Introduction evaluations while minimizing the burden on health-care
providers. The integration of AI algorithms, such as the
Amyotrophic lateral sclerosis (ALS) is a devastating Vivaldi AI system, further augments the capabilities of
neurodegenerative disease, characterized by progressive robotic devices, potentially improving the accuracy and
impairment of motor function, which ultimately leads reliability of functional assessments. 13
to paralysis and respiratory failure. In most cases, ALS
1
manifests with either a spinal-onset disease characterized Despite the potential benefits, the adoption of robotic
by muscle weakness, cramps, and fasciculations, or a bulbar- AI-enabled systems in clinical practice for ALS assessment
onset disease marked by dysarthria and dysphagia. Among requires rigorous validation and exploration of patient
neurodegenerative disorders, ALS is notably the most acceptance and usability. This study aims to address these
rapidly fatal. The progressive deterioration of symptoms gaps by evaluating the feasibility and efficacy of a Sanbot
2
culminates in paralysis and ultimately respiratory failure, Elf robotic device equipped with the Vivaldi AI system
typically resulting in death within an average period of 3 – in administering the ALSFRS-R and assessing patient
5 years from symptom onset. 3 emotional reactions to this novel approach. Through a
comprehensive assessment of functional status and patient
The management of ALS heavily relies on the accurate perspectives, we seek to elucidate the role of robotic
functional assessment, which is used to monitor disease AI technology in advancing healthcare operations and
progression and optimize patient care. At present, management for ALS patients.
one of the most recognized tools for monitoring the
progression of functional impairments in ALS is the ALS 2. Sanbot Elf robot description and
Functional Rating Scale–Revised (ALSFRS-R). It is a architecture
4,5
multidimensional questionnaire of 12 items, which can be
divided into four domains (bulbar, fine and gross motor The Sanbot Elf is a humanoid intelligent service robot,
functions, and respiratory functions). For each question, developed by Qihan Technology. With its bipedal
the score ranges from 0 (complete loss of function) to 4 configuration, it can effortlessly navigate human
(normal function). environments, ensuring seamless interaction through its
ergonomic design. The Sanbot Elf has a total size of 92 × 30
Regulatory agencies recommend using ALSFRS-R in × 40 cm and weighing 19 kg (Figure 1). 13-16 Equipped with
clinical trials, and a study found that the scale was used as a trio of cameras – two within the head and one integrated
the primary endpoint in 82% of therapeutic trials of ALS. into the chest tablet – Sanbot boasts an HD Camera and
6
The ALSFRS-R is easy to administer, clinically meaningful, a 3D Camera with resolutions of 8.0 MP and 1.0 MP,
and sensitive – properties that position it as an efficient respectively. The HD Camera offers functionalities such
tool to assess patients’ disease progression over time as photography, videography, audio recording, and live
and survival. In a 9-month phase III trial of gabapentin, streaming. In addition, Sanbot incorporates a microphone
the scale was proved to wield the most predictive power in its head, three head sensors, and a rear-projector
among the examined outcome measures, such as the boasting a 1920 × 720, 60 Hz, 16:9 resolution.
forced vital capacity and the maximal voluntary isometric
contraction. 7-9 The chest tablet serves as a versatile interface for
displaying and utilizing installed applications, operating on
Despite all these advantages, the scale has some a modified Android OS, and facilitating app development
drawbacks. First, specific training related to the through Android Studio and the Sanbot SDK. Sanbot
administration and to the scoring is required for the proper is further enhanced with a plethora of sensors including
use of ALSFRS-R. Moreover, when a human operator infrared, touch, and passive infrared sensor, alongside
administers the questionnaire, additional information wireless connectivity options such as WIFI, Bluetooth, and
provided by patients and their requests for clarification ZigBee (Figure 1). The core components of the assistive
with respect to unclear questions may be observed in application consist of an Intel NUC7i7-BNH, which
the clinical practice. The combination of these issues features an Intel i7-7567U processor clocked at 3.50 GHz,
could generally result in longer administration times and 16GB DDR4 RAM, and a 256 GB SSD. This setup serves
increased workload for the operators. as both the applications server and the middleware layer.
In recent years, there has been a growing interest in Inter-device communication, including with the robot,
leveraging robotic technology integrated with artificial occurs through a Wi-Fi router. The combination of
intelligence (AI) to enhance functional assessment in the Intel NUC and the robot forms the backbone of the
ALS patients. 10-12 These technological advancements hold assistive application, boasting ample processing power
promise for providing objective, standardized, and efficient and resources to effectively execute all intended tasks

Volume 1 Issue 4 (2024) 74 doi: 10.36922/aih.3732