Retrospective CT/MRI Texture Analysis of Rapidly Progressive Hepatocellular Carcinoma
Rapidly progressive hepatocellular carcinoma (RPHCC) is a subset of hepatocellular carcinoma that demonstrates accelerated growth, and the radiographic features of RPHCC versus non-RPHCC have not been determined. The purpose of this retrospective study was to use baseline radiologic features and tex...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-09-01
|
| Series: | Journal of Personalized Medicine |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2075-4426/10/3/136 |
| _version_ | 1797553117761372160 |
|---|---|
| author | Charissa Kim Natasha Cigarroa Venkateswar Surabhi Balaji Ganeshan Anil K. Pillai |
| author_facet | Charissa Kim Natasha Cigarroa Venkateswar Surabhi Balaji Ganeshan Anil K. Pillai |
| author_sort | Charissa Kim |
| collection | DOAJ |
| description | Rapidly progressive hepatocellular carcinoma (RPHCC) is a subset of hepatocellular carcinoma that demonstrates accelerated growth, and the radiographic features of RPHCC versus non-RPHCC have not been determined. The purpose of this retrospective study was to use baseline radiologic features and texture analysis for the accurate detection of RPHCC and subsequent improvement of clinical outcomes. We conducted a qualitative visual analysis and texture analysis, which selectively extracted and enhanced imaging features of different sizes and intensity variation including mean gray-level intensity (mean), standard deviation (SD), entropy, mean of the positive pixels (MPP), skewness, and kurtosis at each spatial scaling factor (SSF) value of RPHCC and non-RPHCC tumors in a computed tomography (CT) cohort of n = 11 RPHCC and n = 11 non-RPHCC and a magnetic resonance imaging (MRI) cohort of n = 13 RPHCC and n = 10 non-RPHCC. There was a statistically significant difference across visual CT irregular margins <i>p</i> = 0.030 and CT texture features in SSF between RPHCC and non-RPHCC for SSF-6, coarse-texture scale, mean <i>p</i> = 0.023, SD <i>p</i> = 0.053, MPP <i>p</i> = 0.023. A composite score of mean SSF-6 binarized + SD SSF-6 binarized + MPP SSF-6 binarized + irregular margins was significantly different between RPHCC and non-RPHCC (<i>p</i> = 0.001). A composite score ≥3 identified RPHCC with a sensitivity of 81.8% and specificity of 81.8% (AUC = 0.884, <i>p</i> = 0.002). CT coarse-texture-scale features in combination with visually detected irregular margins were able to statistically differentiate between RPHCC and non-RPHCC. By developing an image-based, non-invasive diagnostic criterion, we created a composite score that can identify RPHCC patients at their early stages when they are still eligible for transplantation, improving the clinical course of patient care. |
| first_indexed | 2024-03-10T16:11:47Z |
| format | Article |
| id | doaj.art-9fec94fe3aca40e69be4e125b2cf92e5 |
| institution | Directory Open Access Journal |
| issn | 2075-4426 |
| language | English |
| last_indexed | 2024-03-10T16:11:47Z |
| publishDate | 2020-09-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Journal of Personalized Medicine |
| spelling | doaj.art-9fec94fe3aca40e69be4e125b2cf92e52023-11-20T14:26:58ZengMDPI AGJournal of Personalized Medicine2075-44262020-09-0110313610.3390/jpm10030136Retrospective CT/MRI Texture Analysis of Rapidly Progressive Hepatocellular CarcinomaCharissa Kim0Natasha Cigarroa1Venkateswar Surabhi2Balaji Ganeshan3Anil K. Pillai4Department of Surgery, Huntington Memorial Hospital, 100 W California Blvd, Pasadena, CA 91105, USADepartment of Diagnostic and Interventional Imaging, McGovern Medical School at UTHealth, 6431 Fannin St, Houston, TX 77030, USADepartment of Diagnostic and Interventional Imaging, McGovern Medical School at UTHealth, 6431 Fannin St, Houston, TX 77030, USAInstitute of Nuclear Medicine, University College Medicine, 5th Floor, Tower University College Hospital, 235 Euston Road, London NW1 2BU, UKDivision of Vascular Interventional Radiology, The University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390, USARapidly progressive hepatocellular carcinoma (RPHCC) is a subset of hepatocellular carcinoma that demonstrates accelerated growth, and the radiographic features of RPHCC versus non-RPHCC have not been determined. The purpose of this retrospective study was to use baseline radiologic features and texture analysis for the accurate detection of RPHCC and subsequent improvement of clinical outcomes. We conducted a qualitative visual analysis and texture analysis, which selectively extracted and enhanced imaging features of different sizes and intensity variation including mean gray-level intensity (mean), standard deviation (SD), entropy, mean of the positive pixels (MPP), skewness, and kurtosis at each spatial scaling factor (SSF) value of RPHCC and non-RPHCC tumors in a computed tomography (CT) cohort of n = 11 RPHCC and n = 11 non-RPHCC and a magnetic resonance imaging (MRI) cohort of n = 13 RPHCC and n = 10 non-RPHCC. There was a statistically significant difference across visual CT irregular margins <i>p</i> = 0.030 and CT texture features in SSF between RPHCC and non-RPHCC for SSF-6, coarse-texture scale, mean <i>p</i> = 0.023, SD <i>p</i> = 0.053, MPP <i>p</i> = 0.023. A composite score of mean SSF-6 binarized + SD SSF-6 binarized + MPP SSF-6 binarized + irregular margins was significantly different between RPHCC and non-RPHCC (<i>p</i> = 0.001). A composite score ≥3 identified RPHCC with a sensitivity of 81.8% and specificity of 81.8% (AUC = 0.884, <i>p</i> = 0.002). CT coarse-texture-scale features in combination with visually detected irregular margins were able to statistically differentiate between RPHCC and non-RPHCC. By developing an image-based, non-invasive diagnostic criterion, we created a composite score that can identify RPHCC patients at their early stages when they are still eligible for transplantation, improving the clinical course of patient care.https://www.mdpi.com/2075-4426/10/3/136hepatocellular carcinomafeature detectionqualitative visual analysistexture analysis |
| spellingShingle | Charissa Kim Natasha Cigarroa Venkateswar Surabhi Balaji Ganeshan Anil K. Pillai Retrospective CT/MRI Texture Analysis of Rapidly Progressive Hepatocellular Carcinoma Journal of Personalized Medicine hepatocellular carcinoma feature detection qualitative visual analysis texture analysis |
| title | Retrospective CT/MRI Texture Analysis of Rapidly Progressive Hepatocellular Carcinoma |
| title_full | Retrospective CT/MRI Texture Analysis of Rapidly Progressive Hepatocellular Carcinoma |
| title_fullStr | Retrospective CT/MRI Texture Analysis of Rapidly Progressive Hepatocellular Carcinoma |
| title_full_unstemmed | Retrospective CT/MRI Texture Analysis of Rapidly Progressive Hepatocellular Carcinoma |
| title_short | Retrospective CT/MRI Texture Analysis of Rapidly Progressive Hepatocellular Carcinoma |
| title_sort | retrospective ct mri texture analysis of rapidly progressive hepatocellular carcinoma |
| topic | hepatocellular carcinoma feature detection qualitative visual analysis texture analysis |
| url | https://www.mdpi.com/2075-4426/10/3/136 |
| work_keys_str_mv | AT charissakim retrospectivectmritextureanalysisofrapidlyprogressivehepatocellularcarcinoma AT natashacigarroa retrospectivectmritextureanalysisofrapidlyprogressivehepatocellularcarcinoma AT venkateswarsurabhi retrospectivectmritextureanalysisofrapidlyprogressivehepatocellularcarcinoma AT balajiganeshan retrospectivectmritextureanalysisofrapidlyprogressivehepatocellularcarcinoma AT anilkpillai retrospectivectmritextureanalysisofrapidlyprogressivehepatocellularcarcinoma |