Electrodiagnostic accuracy in polyneuropathies: supervised learning algorithms as a tool for practitioners

Machine Learning for Early Diagnosis of ATTRv Amyloidosis in Non-Endemic Areas: A Multicenter Study from Italy

BACKGROUND

Hereditary transthyretin amyloidosis with polyneuropathy (ATTRv) is an adult-onset multisystemic disease, affecting the peripheral nerves, heart, gastrointestinal tract, eyes, and kidneys. Nowadays, several treatment options are available; thus, avoiding misdiagnosis is crucial to starting therapy in early disease stages. However, clinical diagnosis may be difficult, as the disease may present with unspecific symptoms and signs. We hypothesize that the diagnostic process may benefit from the use of machine learning (ML).

METHODS

397 patients referring to neuromuscular clinics in 4 centers from the south of Italy with neuropathy and at least 1 more red flag, as well as undergoing genetic testing for ATTRv, were considered. Then, only probands were considered for analysis. Hence, a cohort of 184 patients, 93 with positive and 91 (age- and sex-matched) with negative genetics, was considered for the classification task. The XGBoost (XGB) algorithm was trained to classify positive and negative TTR mutation patients. The SHAP method was used as an explainable artificial intelligence algorithm to interpret the model findings.

RESULTS

diabetes, gender, unexplained weight loss, cardiomyopathy, bilateral carpal tunnel syndrome (CTS), ocular symptoms, autonomic symptoms, ataxia, renal dysfunction, lumbar canal stenosis, and history of autoimmunity were used for the model training. The XGB model showed an accuracy of 0.707 ± 0.101, a sensitivity of 0.712 ± 0.147, a specificity of 0.704 ± 0.150, and an AUC-ROC of 0.752 ± 0.107. Using the SHAP explanation, it was confirmed that unexplained weight loss, gastrointestinal symptoms, and cardiomyopathy showed a significant association with the genetic diagnosis of ATTRv, while bilateral CTS, diabetes, autoimmunity, and ocular and renal involvement were associated with a negative genetic test.

CONCLUSIONS

Our data show that ML might potentially be a useful instrument to identify patients with neuropathy that should undergo genetic testing for ATTRv. Unexplained weight loss and cardiomyopathy are relevant red flags in ATTRv in the south of Italy. Further studies are needed to confirm these findings.

Motor Conduction Studies and Handgrip in Hereditary TTR Amyloidosis: Simple Tools to Evaluate the Upper Limbs

Purpose

Hereditary transthyretin amyloidosis with polyneuropathy (ATTRv) is caused by mutations in the TTR gene, leading to misfolded monomers that aggregate generating amyloid fibrils. The clinical phenotype is heterogeneous, and characterized by a multisystemic disease affecting the sensorimotor and autonomic functions along with other organs.

Materials and Methods

All the patients were assessed by complete neurological assessment, neurophysiological evaluation, of the median nerve, and handgrip analysis. The data are presented as means and standard deviations. Parametric and non-parametric assessments have been performed to identify differences between groups. Pearson's correlation has been carried out when appropriate.

Results

Twenty patients with ATTRv (66.1 ± 8.4 years; eight females) and 30 controls (61.1 ± 11.6 years; 16 females) were enrolled. Handgrip strength was reduced in patients with ATTR in both right and left hands compared to the controls. Significant differences were found between patients and controls in the right (handgrip right, HGS_R TTR 21.1 ± 13 kg vs. HGS_R Control 29.4 ± 11.3 kg, p = 0.017) and left (handgrip left, HGS_L TTR 22.2 ± 10.7 kg. vs. HGS_L Control 31 ± 11.3 kg, p = 0.007). NIS and CMAP amplitude of the median nerve were related to HGS measures for both hands in patients with ATTRv.

Conclusions

The progression of bilateral carpal tunnel syndrome is related to neurophysiological data in the median nerve in ATTRv. Also, handgrip measures might represent an important tool for the assessment of disease progression in ATTRv. We propose using a combination of CMAP amplitude and HGS for the assessment of hand motor strength in ATTRv.

Evidence base for investigative and therapeutic modalities in chronic inflammatory demyelinating polyneuropathy and multifocal motor neuropathy

Chronic inflammatory demyelinating polyneuropathy, its variants and multifocal motor neuropathy belong to a spectrum of peripheral nerve disorders with complex dysimmune disease mechanisms. Awareness of the unique clinical phenotypes but also heterogeneity between patients is vital to arrive at early suspicion and ordering appropriate tests. This includes requirements for optimal electrodiagnostic protocol, aimed to capture sufficient electrophysiologic evidence for relevant abnormalities, a case-based approach on the eventual need to further expand the diagnostic armamentarium and correct reading of their results. Considerable phenotypical variation, diverse combinations of abnormalities found on diagnostic tests and heterogeneity in disease course and treatment response, all contribute to widespread differences in success rates on timely diagnosis and optimal treatment. We aim to provide a practical overview and guidance on relevant diagnostic and management strategies, including pitfalls and present a summary of the relevant novel developments in this field.

The neurophysiological lesson from the Italian CIDP database

INTRODUCTION

Electrophysiological diagnosis of chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) may be challenging. Thus, with the aim ofproviding some practical advice in electrophysiological approach to a patient with suspected CIDP, we analyzed electrophysiological data from 499 patients enrolled inthe Italian CIDP Database.

METHODS

We calculated the rate of each demyelinating feature, the rate of demyelinating features per nerve, the diagnostic rate for upper andlower limb nerves, and, using a ROC curve analysis, the diagnostic accuracy of each couple of nerves and each demyelinating feature, for every CIDP subtype.Moreover, we compared the electrophysiological data of definite and probable CIDP patients with those of possible and not-fulfilling CIDP patients, and by a logisticregression analysis, we estimated the odds ratio (OR) to make an electrophysiological diagnosis of definite or probable CIDP.

RESULTS

The ulnar nerve had the highestrate of demyelinating features and, when tested bilaterally, had the highest diagnostic accuracy except for DADS in which peroneal nerves were the most informative.In possible and not-fulfilling CIDP patients, a lower number of nerves and proximal temporal dispersion (TD) measurements had been performed compared to definiteand probable CIDP patients. Importantly, OR for each tested motor nerve and each TD measurement was 1.59 and 1.33, respectively.

CONCLUSION

Our findingsdemonstrated that the diagnosis of CIDP may be missed due to inadequate or incomplete electrophysiological examination or interpretation. At the same time, thesedata taken together could be useful to draw a thoughtful electrophysiological approach to patients suspected of CIDP.

The neurophysiological lesson from the Italian CIDP database

Introduction

Electrophysiological diagnosis of chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) may be challenging. Thus, with the aim ofproviding some practical advice in electrophysiological approach to a patient with suspected CIDP, we analyzed electrophysiological data from 499 patients enrolled inthe Italian CIDP Database.

Methods

We calculated the rate of each demyelinating feature, the rate of demyelinating features per nerve, the diagnostic rate for upper andlower limb nerves, and, using a ROC curve analysis, the diagnostic accuracy of each couple of nerves and each demyelinating feature, for every CIDP subtype.Moreover, we compared the electrophysiological data of definite and probable CIDP patients with those of possible and not-fulfilling CIDP patients, and by a logisticregression analysis, we estimated the odds ratio (OR) to make an electrophysiological diagnosis of definite or probable CIDP.

Results

The ulnar nerve had the highestrate of demyelinating features and, when tested bilaterally, had the highest diagnostic accuracy except for DADS in which peroneal nerves were the most informative.In possible and not-fulfilling CIDP patients, a lower number of nerves and proximal temporal dispersion (TD) measurements had been performed compared to definiteand probable CIDP patients. Importantly, OR for each tested motor nerve and each TD measurement was 1.59 and 1.33, respectively.

Conclusion

Our findingsdemonstrated that the diagnosis of CIDP may be missed due to inadequate or incomplete electrophysiological examination or interpretation. At the same time, thesedata taken together could be useful to draw a thoughtful electrophysiological approach to patients suspected of CIDP.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10072-021-05321-z.