Home > Riviste > Journal of Radiological Review > Fascicoli precedenti > Journal of Radiological Review 2021 June;8(2) > Journal of Radiological Review 2021 June;8(2):152-7



Opzioni di pubblicazione
Per abbonarsi
Segnala alla tua biblioteca


Publication history
Per citare questo articolo



Journal of Radiological Review 2021 June;8(2):152-7

DOI: 10.23736/S2723-9284.21.00127-9


lingua: Inglese

Role of quantitative imaging and deep learning in interstitial lung diseases

Salvatore C. FANNI 1, Caterina A. D’AMORE 1 , Alessio MILAZZO 1, Annalisa DE LIPERI 2, Lucio CALANDRIELLO 3, Giuseppe CICCHETTI 3, Elisa BARATELLA 4, Adele VALENTINI 5, Anna Rita LARICI 3, 6, Chiara ROMEI 2

1 Unit of Diagnostic Radiology 1, Department of Translational Research, Pisa University Hospital, Pisa, Italy; 2 Unit of Diagnostic Radiology 2, Department of Diagnostic Imaging, Pisa University Hospital, Pisa, Italy; 3 Unit of Diagnostic Imaging Area, Department of Diagnostic Imaging, Oncological Radiotherapy and Hematology, “A. Gemelli” University Polyclinic Foundation IRCCS, Rome, Italy; 4 Department of Radiology, University of Trieste, Hospital of Cattinara, Trieste, Italy; 5 Department of Radiology, IRCCS San Matteo Polyclinic Foundation, Pavia, Italy; 6 Department of Radiological and Hematological Sciences, Section of Radiology, Catholic University of the Sacred Heart of Rome, Rome, Italy

Interstitial lung disease (ILD) are a large group of diffuse lung diseases characterized by similar clinical, pathological and radiological features. High resolution computed tomography (HRCT) has a central role in ILD diagnosis and management. In the last few years, computer-aided methods as Quantitative Computer Tomography (QCT) and Artificial Intelligence (AI) software were proposed as a source of reliable quantitative imaging biomarkers. The present review aimed to summarize and describe the current QCT and AI methods and to evaluate their potential diagnostic and prognostic role. The first attempt to a quantitative analysis of HRCT in ILD is represented by the density histogram analysis with the definition of two new parameter, Kurtosis and Skewness. Then texture analysis tools were developed as Adaptive Multiple Features Method (AMFM), Computer-Aided Lung Informatics for Pathology Evaluation and Ratings (CALIPER), Quantitative Lung Fibrosis (QLF) and Automated Quantification System (AQS). The introduction of AI technologies further increased the amount of objective and reproducible biomarkers. The diagnostic and prognostic role of QCT and AI methods was analyzed and confirmed in various studies, as reported in the review. QCT and AI technologies application led to the introduction of new objective biomarkers with relevant diagnostic and prognostic implications. However, there is still the need for more prospective study and the creation of open-source datasets would help to assess QCT and AI methods efficacy and to compare them.

KEY WORDS: Lung diseases, interstitial; Tomography, X-ray computed; Artificial intelligence; Machine learning; Idiopathic pulmonary fibrosis

inizio pagina