Machine learning-based FDG PET-CT radiomics for outcome prediction in larynx and hypopharynx squamous cell carcinoma

Published on Jan 1, 2021in Clinical Radiology2.118
· DOI :10.1016/J.CRAD.2020.08.030
Jim Zhong5
Estimated H-index: 5
(Leeds Teaching Hospitals NHS Trust),
R. Frood6
Estimated H-index: 6
(Leeds Teaching Hospitals NHS Trust)
+ 6 AuthorsAndrew Scarsbrook29
Estimated H-index: 29
(Leeds Teaching Hospitals NHS Trust)
AIM To determine whether machine learning-based radiomic feature analysis of baseline integrated 2-[18F]-fluoro-2-deoxy- d- glucose (FDG) positron-emission tomography (PET) computed tomography (CT) predicts disease progression in patients with locally advanced larynx and hypopharynx squamous cell carcinoma (SCC) receiving (chemo)radiotherapy. MATERIALS AND METHODS Patients with larynx and hypopharynx SCC treated with definitive (chemo)radiotherapy at a specialist cancer centre undergoing pre-treatment PET-CT between 2008 and 2017 were included. Tumour segmentation and radiomic analysis was performed using LIFEx software (University of Paris-Saclay, France). Data were assigned into training (80%) and validation (20%) cohorts adhering to TRIPOD guidelines. A random forest classifier was created for four predictive models using features determined by recursive feature elimination: (A) PET, (B) CT, (C) clinical, and (D) combined PET-CT parameters. Model performance was assessed using area under the curve (AUC) receiver operating characteristic (ROC) analysis. RESULTS Seventy-two patients (40 hypopharynx 32 larynx tumours) were included, mean age 61 (range 41–77) years, 50 (69%) were men. Forty-five (62.5%) had chemoradiotherapy, 27 (37.5%) had radiotherapy alone. Median follow-up 26 months (range 12–105 months). Twenty-seven (37.5%) patients progressed within 12 months. ROC AUC for models A, B, C, and D were 0.91, 0.94, 0.88, and 0.93 in training and 0.82, 0.72, 0.70, and 0.94 in validation cohorts. Parameters in model D were metabolic tumour volume (MTV), maximum CT value, minimum standardized uptake value (SUVmin), grey-level zone length matrix (GLZLM) small-zone low grey-level emphasis (SZLGE) and histogram kurtosis. CONCLUSION FDG PET-CT derived radiomic features are potential predictors of early disease progression in patients with locally advanced larynx and hypopharynx SCC.
📖 Papers frequently viewed together
75 Citations
25 Citations
60 Citations
#1Xiaokai Mo (STU: Shantou University)H-Index: 9
#1Xiaokai Mo (Peking Union Medical College)H-Index: 10
Last. Shuixing Zhang (JNU: Jinan University)H-Index: 18
view all 16 authors...
Purpose To develop a radiomics-based model to stratify the risk of early progression (local/regional recurrence or metastasis) among patients with hypopharyngeal cancer undergoing chemoradiotherapy and modify their pretreatment plans.
12 CitationsSource
#2Oier Echaniz (UPV/EHU: University of the Basque Country)H-Index: 3
Last. Manuel Graña (UPV/EHU: University of the Basque Country)H-Index: 31
view all 6 authors...
Radiomics and texture analysis represent a new option in our biomarkers arsenal. These techniques extract a large number of quantitative features, analyzing their properties to incorporate them in clinical decision-making. Laryngeal cancer represents one of the most frequent cancers in the head and neck area. We hypothesized that radiomics features can be included as a laryngeal cancer precision medicine tool, as it is able to non-invasively characterize the overall tumor accounting for heteroge...
8 CitationsSource
#1M. ArunsinghH-Index: 5
#2Sriram VaidyanathanH-Index: 15
Last. Robin PrestwichH-Index: 28
view all 6 authors...
Abstract Aim There are few data to inform on the use of response assessment 2-[fluorine-18]-fluoro-2-deoxy- d -glucose (FDG) positron emission tomography-computed tomography (PET-CT) following radical radiotherapy without chemotherapy for head and neck squamous cell carcinoma (HNSCC). This retrospective study evaluated the accuracy of PET-CT in HNSCC following radical radiotherapy. Materials and methods In total, 138 patients with HNSCC treated with radical radiotherapy without chemotherapy who ...
7 CitationsSource
#1Paul Giraud (Paris V: Paris Descartes University)H-Index: 3
#2Philippe Giraud (Paris V: Paris Descartes University)H-Index: 29
Last. Jean-Emmanuel Bibault (Paris V: Paris Descartes University)H-Index: 19
view all 10 authors...
Introduction: An increasing number of parameters can be considered when making decisions in oncology. Tumor characteristics can also be extracted from imaging through the use of radiomics and add to this wealth of clinical data. Machine learning can encompass these parameters and thus enhance clinical decision as well as radiotherapy workflow. Methods: We performed a description of machine learning applications at each step of treatment by radiotherapy in head and neck cancers. We then performed...
39 CitationsSource
: Patients with hypopharyngeal cancers frequently present at advanced stage and in poor general health status. Their natural history is characterised by diffuse local primary disease, mucosal and submucosal spread, early cervical nodal metastasis, and a relatively high rate of distant spread. By the time of initial diagnosis, some 60% of all hypopharyngeal cancer patients will be with stage IV disease, some 5% will present with distant metastases, and almost 40% will have a significant reduction...
3 CitationsSource
BACKGROUND: This study aimed at assessing the prognostic value of textural indices extracted from 18F-fluorodeoxyglucose positron-emission tomography (FDG-PET)/CT in a large cohort of patients with head and neck squamous cell carcinomas (HNSCC) of any anatomic subsite and staging. METHODS: Consecutive patients with HNSCC referred for a pretreatment FDG-PET/CT were retrospectively included and followed up for a minimum of 2 years. Standardized uptake value, metabolic tumor volume (MTV), and textu...
13 CitationsSource
#1Thong Chotchutipan (UM: University of Michigan)H-Index: 2
#2B.S. Rosen (UM: University of Michigan)H-Index: 7
Last. Michelle Mierzwa (UM: University of Michigan)H-Index: 15
view all 9 authors...
BACKGROUND: We sought to investigate the prognostic value of volumetric positron emission tomography (PET) parameters in patients with human papillomavirus (HPV)-related oropharyngeal squamous cell carcinoma (OPSCC) and a ≤10 pack-year smoking history treated with chemoradiation. METHODS: A total of 142 patients were included. Maximum standardized uptake value, metabolic tumor volume, and total lesion glycolysis (TLG) of the primary tumor, involved regional lymph nodes, and total lesion were cal...
10 CitationsSource
#1Christophe Nioche (Université Paris-Saclay)H-Index: 15
#2Fanny Orlhac (Université Paris-Saclay)H-Index: 16
Last. Irène Buvat (Université Paris-Saclay)H-Index: 46
view all 10 authors...
Textural and shape analysis is gaining considerable interest in medical imaging, particularly to identify parameters characterizing tumor heterogeneity and to feed radiomic models. Here, we present a free, multiplatform, and easy-to-use freeware called LIFEx, which enables the calculation of conventional, histogram-based, textural, and shape features from PET, SPECT, MR, CT, and US images, or from any combination of imaging modalities. The application does not require any programming skills and ...
254 CitationsSource
#1Antoine Schernberg (Institut Gustave Roussy)H-Index: 1
#1Antoine Schernberg (French Institute of Health and Medical Research)H-Index: 1
Last. Charlotte RobertH-Index: 17
view all 12 authors...
Purpose We investigated whether a score combining baseline neutrophilia and a PET biomarker could predict outcome in patients with locally advanced cervical cancer (LACC).
21 CitationsSource
#1Kieran Foley (Cardiff University)H-Index: 9
#2Robert Kerrin Hills (Cardiff University)H-Index: 76
Last. Stuart A. G. Roberts (University Hospital of Wales)H-Index: 23
view all 9 authors...
Objectives This retrospective cohort study developed a prognostic model incorporating PET texture analysis in patients with oesophageal cancer (OC). Internal validation of the model was performed. Methods Consecutive OC patients (n = 403) were chronologically separated into development (n = 302, September 2010-September 2014, median age = 67.0, males = 227, adenocarcinomas = 237) and validation cohorts (n = 101, September 2014-July 2015, median age = 69.0, males = 78, adenocarcinomas = 79). Text...
34 CitationsSource
Cited By0