RAN-IO: Radiomics as biomArker imaging for NSCLC patients treated with ImmunOtherapy

Dr. Arsela Prelaj (arsela.prelaj@polimi.it)

Aldo Marzullo  (aldo.marzullo@polimi.it)

Aim:

Predicting Immunotherapy (IO) efficacy in Non-Small Cell Lung Cancer (NSCLC) is a crucial unmet clinical need. Beyond Programmed Death Ligand-1 (PD-L1) no other biomarkers are approved. However, PD-L1 is quite Imperfect. Radiomics was recently used as imaging biomarker in NSCLC patients treated with IO. This thesis aim is to find a predictive radiomic signature for NSCLC patients treated with IO

Project phases:

• Literature review

• Development of  predictive radiomic signature

Requirements:

• Knowledge of Matlab
• Knowledge of Python (scikit-learn, TensorFlow/Keras)

• Basic knowledge of Git and Linux is welcome

Dr. Arsela Prelaj (arsela.prelaj@polimi.it)

Dr. Mattia Pesenti (mattia.pesenti@polimi.it)

Aim:

Osimertinib (OSI), a third-generation tyrosine kinase inhibitor (TKI) is now become the standard of care and is given as first line therapy in NSCLC patients with EGFR mutation (mut) since it demonstrated superiority compare to 1st and 2nd generation (G) TKI. Previously, OSI was received in 2nd line  after disease progression upon 1st or 2nd  generation TKI in around 50% of patients who acquired an exon20 resistance mutation called T790M. Recent data demonstrated that combination therapy with 1st G TKI in first Line obtained the same PFS compare to OSI. Thus predicting the onset of T790M can prolong survival of around 12 months in around 50% using combo therapy in first line and OSI in second line as sequence therapy. The aim of this study is to predict T790M using AI techniques by integrating all data of NSCLC patients with EGFR mut (Fig in the right).

Project phases:

• Literature review
• Development of  AI software to predict T790M

Requirements:

• Knowledge of Matlab
• Knowledge of Python (scikit-learn, TensorFlow/Keras)

• Basic knowledge of Git and Linux is welcome

Dr. Arsela Prelaj (arsela.prelaj@polimi.it)

Dr. Mattia Pesenti (mattia.pesenti@polimi.it)

Aim:

Predicting Immunotherapy (IO) efficacy oin NSCLC is a crucial unmet clinical need. Beyond Programmed Death Ligand-1 (PD-L1) no other biomarkers are approved. However, PD-L1 is quite Imperfect. Thus, AI which is able to integrate high dimensional real world and multiomics data could be a unique approach able To find an algorithm able to predict IO in NSCLC patients. The later will be this thesis aim.

Project phases:

• Literature review
• Development of  AI software to predict IO inNSCLC

Requirements:

• Knowledge of Matlab
• Knowledge of Python (scikit-learn, TensorFlow/Keras)

• Basic knowledge of Git and Linux is welcome

Aim:

Essential tremor (ET) is the most common movement disorder, affecting an estimated 3% of the population. Focused ultrasound is a noninvasive, therapeutic technology with the potential to improve the quality of life and decrease the cost of care for patients with essential tremor. This novel technology focuses multiple beams of ultrasound energy precisely and accurately on targets deep in the brain without damaging surrounding normal tissue.

Pre- and post- treatment assessment are conducted on a cohort of patients recruited at the Besta Institute. Evaluation are conducted through clinical scales and through an instrumented smart ink pen able to extract relevant clinical parameters.

Project phases:​

• Literature review
• Data analysis and algorithm development
• Statistics (pre-post treatment effect; correlation with clinical scales; prediction of treatment outcome)

Requirements:​

• Basic knowledge of Matlab
• Interest in statistics and data analysis

Aim:

Premature babies need a special nutrition program to overcome their frailty condition. It must balance different nutrients apport, to achieve a target growth. However, the real effect of different nutrition programs is still unclear. The aim of this work is the creation of a machine learning model on babies’ response to nutritional programs and the impact on their global health.

Project phases:​

• Literature review
• Machine Learning models implementation

Requirements:​

• Interest in machine learning

Aim:

Gait impairments, including bradykinesia and freezing of gait (FoG), are the most common and disabling symptoms in Parkinson’s disease patients. Quantifying gait impairments under free-living and challenging condition (e.g. FoG-provoking test) using sensing technologies is a promising avenue to assess and monitor disease severity. The aim of this work is to use sensing technologies to quantify gait impairment in Parkinson’s disease, with the final aim of defining tailored interventions.

Project phases:​

• Literature review​
• Set-up and protocol refinement
• Pilot testing on patients with Parkinson’s disease
• Data analysis

​Requirements:​

• Interest in working with patients
• Basic knowledge of Matlab and programming languages

Aim:

Specific Learning Disabilities (SLDs) first screening starts from direct teachers’ observation. If it is not possible, e.g., in distance learning during COVID-19 lockdown, they exacerbate and prevent children from proper learning. The aim of this work is to provide serious games which children would like to play at home and could provide hints of difficulties in reading, writing or calculation.

Project phases:

• Literature review
• Exergame design and implementation
• Preliminary tests on target users

Requirements:

• Interest in game development
• Availability to learn programming languages

Aim:

Dementia affects millions of people worldwide. Unfortunately, it cannot be cured, but an early diagnosis can help to better manage the disease evolution. Handwriting results from a complex network made up of cognitive, kinesthetic, and perceptual-motor abilities and it is one of the daily’s activities affected in patients with dementia. 

The aim of this work is to use an IoT smart ink pen for the characterization of patients with dementia and mild cognitive impairment with the final goal of supporting early diagnosis. 

Project phases:

• Literature review
• Data collection
• Data analysis and algorithm development

Requirements:

• Basic knowledge of Matlab
• Basic knowledge of Python

Aim:

The diagnosis of clinically probable Parkinson’s Disease (PD) in the early stages relies primarily on clinical assessment by a neurologist. When motor symptoms affect the dominant hand, patients may report worsening of

handwriting as of the initial symptoms. The aim of this work is to use an IoT smart ink pen for the characterization of PD patients handwriting, with the final goal of supporting PD patients’ diagnosis and remote monitoring

Project phases:​

• Literature review​
• Data collection
• Data analysis and algorithm development

​Requirements:​

• Basic knowledge of Matlab
• Basic knowledge of Python

Aim:

Handwriting is a high-value task entailing a unique blend of cognitive, perceptual, and fine motor skills and for this reason its assessment is leveraged in a number of health-related applications, including the diagnostic process. The aim of this work is to profile normal handwriting and potential deviations through data longitudinally collected by an IoT smart ink pen and anomaly detection techniques.

Project phases:

• Literature review
• Data collection
• Data analysis ald algorithm development

Requirements:

• Basic knowledge of Matlab or Python

Aim:

Handwriting is a high-value task entailing a unique blend of cognitive, perceptual, and fine motor skills and for this reason its assessment is leveraged in a number of health-related applications,  One of the aim of the European ESSENCE project is to leverage a novel IoT smart in pen to monitor user’s handwriting. 

The aim of this thesis work is to develop a companion interface to manage – via Bluetooth – communication, data recording, data download and important  functionalities related to the pen use.

Project phases:

• Research for identification of solutions​
• Companion Interface design of automated algorithms
• Companion interfacedevelopment
• Companion Interface testing

Requirements:

• KnowledgeAndroid programming or otherlanguages (Python, Javascript…)
• Interest in programming and problem solving

Aim:

Selecting and evaluating appropriate treatment for children with hypertonic movement disorders is nontrivial. One challenge is the ability of quantifying the presence and importance of motor impairments, especially when more than one coexist. This is the case, for instance, of mixed hypertonia with components of spasticity and dystonia. Against this background, the DYSPA System aims at achieving quantitative assessment, encompassing kinematic and electromyographic measures, that quantifies neuro-motor performance during functional tasks and measures the presence and extent of motor impairments through specific dystonia and spasticity indices.

Project phases:

• Literature review
• Data collection on children with movement disorders and age-matched controls
• Data analysis and statistics

Requirements:

• Knowledge of Matlab
• Interest in learning other analysis softwares (e.g., BTS Smart Analyzer, R)
• Interest in pediatric movement disorders
• Interest in working in close contact with patients

Aim:

Speech conveys important information about health condition and quality of life. Voice parameters can inform about cognitive decline, stress, emotions and arousal. Therefore, automatic voice analysis is a good candidate to be included in mobile-health technologies, offering an ecological and continuous monitoring. Within a previous European project (MoveCare), a mobile app was developed to identify cognitive decline based on voice parameters computed on the fly during phone conversations. This work, included in the European project ESSENCE, aims at extending this app including features correlated also to emotional valence and arousal.

Project phases:

• Literature review
• Implementation of an algorithm to automatically derive voice features
• Machine Learning models implementation
• Design of a mobile app

Requirements:

• Knowledge of Matlab
• Interest in machine learning and app development

Aim:

In early school years, poor handwriting development has negative consequences on children self esteem and behaviour. Monitoring subtle changes is important, but they cannot be disclose by pure observation only.
The aim of this work is to profile the development of children’s normal handwriting and potential deviations through data longitudinally collected by an IoT smart ink pen and anomaly detection techniques.

Project phases:

• Literature review
• Data collection
• Data analysis

Requirements:

• Basic knowledge of Matlab or Python

Aim:

An inefficient grasping position causes difficulties in handwriting and an early correction can avoid persistent problems. In distance learning, teachers’ or experts’ direct observation was not possible.
The aim of this work is to leverage computer vision to detect the grasping strategy of children, through video recordings of handwriting production, and to relate it to pen movements collected by an IoT smart ink pen, with the final goal of understanding grasping problems from IoT sensors only.

Project phases:

• Literature review
• Data collection
• Computer vision algorithms implementation
• IoT sensors’ data analysis

Requirements:

• Basic knowledge of computer vision and deep learning
• Availability to reach schools for data collection (province of Varese)

Aim:

Voice signals provide important information to measure human behaviors and cognitive functions. Various types of neurodegenerative dementia (Alzheimer’s disease, frontotemporal dementia, primary progressive aphasia, Lewy body dementia) affect human speech in different manners. Disorders or irregularities in language domain, evaluated in terms of temporal features, such as pauses, speech rate, pitch, etc, could be strong predictors of different neurodegenerative pathologies affecting the brain. This thesis aims at the development and validation of a software tool able to automatically extract a set of language features sensitive and to classify neurodegenerative disorders from voice signals.

Project phases:

Starting from a Matlab-algorithm to automatically extract voice features from recordings and a dataset of recordings on healthy elderly subjects:

– Literature review on vocal indicators of cognitive decline
– Creation of a database of normal and pathological voice samples from elderly (control subjects, patients with Alzheimer’s disease, frontotemporal dementia, primary progressive aphasia and Lewy body dementia)
– Optimization of the algorithm to extract voice features
– Development of a classification algorithm to support the diagnosis of neurodegenerative diseases based on voice features and neurodegenerative biomarkers (magnetic resonance atrophy, cerebrospinal fluid proteins concentration)

Requirements:

• Knowledge of Matlab
• Knowledge (or intention to know) Python
• Availability to participate to data collection

Aim:

Deficits in the upper extremity, especially in the hand usually affects motor control ability, leading to difficulties in performing the daily life activities.

This work aims at developing a digital solution for hand rehabilitation using serious games controlled through a sensorized smart ball to aid patients and older adults in rehabilitating their hand.

Project phases:

– Literature review
– Design of serious games
– Tests on healthy subjects and patients
– Data analysis

Requirements:

• Interest in game development
• Availability to learn programming languages

Aim:

An ICT system, encompassing inertial, temperature, electrodermal activity, and EMG sensors has been developed to be used in combination with cognitive assessment to achieve longitudinal monitoring of patients with Disorders of Consciousness (DOC), with the final aim of testing the efficacy of tailored interventions.

Project phases:

– Literature review
-Data acquisition on patients
-Data analysis and design of algorithms to estimate pain and detect voluntary/unvoluntary contractions during long term acquisitions

Aim:

Studies show that the use of systematic observation of meaningful actions followed by their execution (action observation treatment [AOT]) may become a rehabilitative strategy to accelerate the process of functional recovery in patients with motor impairment.

This work aims at developing a home-based system for neurological patients that leverages eye tracking technology during the observation of video clips showing appropriate actions, with the aim of evaluating compliance to the required task. 

Project phases:

– Literature review
-Design of technological solution with definition of requirements and specifications
–Integration of eye tracking with tablet/laptop
–Data acquisition and analysis

Aim:

Dysgraphia diagnosis is mainly based on writing speed on paper, but the causes which underlie such slowness are rarely investigated in clinical practice. Besides the evaluation of the overall speed, the digitalization of a Dysgraphia test would provide additional parameters related to gesture production, such as, fluidity, pressure, and tremor, and their variability during the execution.

The aim of this work is to validate the digital version of a test for Dysgraphia, to develop an end-to-end expert system which automates the diagnosis, and to leverage Machine Learning techniques to provide additional insights on gesture execution, towards a more targeted diagnosis.

Project phases:

– Literature review
– Data collection on healthy and dysgraphic subjects
– Data analysis to extract the score and the features of interest from the acquired data
– Statistical analysis
– Development of Machine Learning models

Requirements:

• MATLAB, R or Python knowledge
• Interest in Machine Learning and Deep Learning
• Availability to reach schools for data acquisition

Aim:

Disentangling transient handwriting difficulties from Dysgraphia is not a trivial task. To facilitate the process, an observational and empowerment study started two years ago.

The aim of this work is to longitudinally monitor handwriting-related problems, starting from preschoolers, to leverage Machine Learning techniques to predict the level of risk and evaluate the effectiveness of interventions, towards an early screening of Dysgraphia.

Project phases:

– Literature review
– Data collection
– Data Analysis
– Machine Learning algorithms implementation

Requirements:

• MATLAB, R, or Python knowledge
• Interest in machine learning
• Availability to reach schools for data acquisition (province of Varese)