Radiological evaluation of interstitial lung disease:

Lung MRI to predict response or lack of response to treatment in interstitial lung disease: initial observations on SSFSE/PROPELLER T2 match/mismatch

BACKGROUND

Ground-glass opacities (GGO) are frequently found in interstitial lung diseases (ILD) and may represent either active inflammation or subresolution interstitial fibrosis. We sought to investigate the ability of lung MRI to predict treatment response in individuals with ILD presenting with predominant GGO. Methods: prospective cohort, 15 participants presenting with ILD manifested as predominant GGO and referred for a new treatment regimen with a systemic glucocorticoid and/or an immunosuppressive agent, underwent 1.5 T lung MRI. SSFSE/PROPELLER T2 mismatch sign, relative signal intensity on T2-weighted images and relative enhancement of lung lesions were compared to functional response, defined as a greater than 10% increase in forced vital capacity in 10 weeks (primary endpoint).

RESULTS

SSFSE/PROPELLER T2 match/mismatch was able to discriminate responders from nonresponders for the primary endpoint in 12 of 15 participants (80% accuracy, p = 0.026) for readers 1 and 2, and in 13 of 15 participants (87% accuracy, p = 0.011) for reader 3, with interrater agreement of 87% between readers 1 and 2 (Cohen's kappa coefficient of 0.732) and 93% between readers 1/2 and 3 (Cohen's kappa coefficient of 0.865).

CONCLUSIONS

SSFSE-PROPELLER T2 match/mismatch was predictive of treatment response status in this group of ILD patients. Abbreviations FVC: forced vital capacityGGO; ground-glass opacities; HRCT: High-Resolution Computed Tomography; ILD: interstitial lung disease; LAVA: Liver Acquisition with Volume Acceleration; mMRC: modified Medical Research Council dyspnea score; MRI: Magnetic Resonance Image; PROPELLER: Periodically Rotated Overlapping Parallel Lines with Enhanced Reconstruction; SI: signal intensity; SSFSE: Single-Shot Fast Spin Echo.

Exploratory analysis of methods for automated classification of laboratory test orders into syndromic groups in veterinary medicine

BACKGROUND

Recent focus on earlier detection of pathogen introduction in human and animal populations has led to the development of surveillance systems based on automated monitoring of health data. Real- or near real-time monitoring of pre-diagnostic data requires automated classification of records into syndromes--syndromic surveillance--using algorithms that incorporate medical knowledge in a reliable and efficient way, while remaining comprehensible to end users.

METHODS

This paper describes the application of two of machine learning (Naïve Bayes and Decision Trees) and rule-based methods to extract syndromic information from laboratory test requests submitted to a veterinary diagnostic laboratory.

RESULTS

High performance (F1-macro = 0.9995) was achieved through the use of a rule-based syndrome classifier, based on rule induction followed by manual modification during the construction phase, which also resulted in clear interpretability of the resulting classification process. An unmodified rule induction algorithm achieved an F(1-micro) score of 0.979 though this fell to 0.677 when performance for individual classes was averaged in an unweighted manner (F(1-macro)), due to the fact that the algorithm failed to learn 3 of the 16 classes from the training set. Decision Trees showed equal interpretability to the rule-based approaches, but achieved an F(1-micro) score of 0.923 (falling to 0.311 when classes are given equal weight). A Naïve Bayes classifier learned all classes and achieved high performance (F(1-micro)= 0.994 and F(1-macro) = .955), however the classification process is not transparent to the domain experts.

CONCLUSION

The use of a manually customised rule set allowed for the development of a system for classification of laboratory tests into syndromic groups with very high performance, and high interpretability by the domain experts. Further research is required to develop internal validation rules in order to establish automated methods to update model rules without user input.