Comparison of diagnostic accuracy between [18F]FDG PET/MRI and contrast-enhanced MRI in T staging for oral tongue cancer

Positron Emission Tomography-Magnetic Resonance Imaging, a New Hybrid Imaging Modality for Dentomaxillofacial Malignancies-A Systematic Review

Background/Objectives: Emerging hybrid imaging modalities, like Positron Emission Tomography/Computed Tomography (PET/CT) and Positron Emission Tomography/Magnetic Resonance Imaging (PET/MRI), are useful for assessing head and neck cancer (HNC) and its prognosis during follow-up. PET/MRI systems enable simultaneous PET and MRI scans within a single session. These combined PET/MRI scanners merge MRI's better soft tissue contrast and the molecular metabolic information offered by PET. Aim: To review scientific articles on the use of hybrid PET/MRI techniques in diagnosing dentomaxillofacial malignancies. Method: The available literature on the use of PET/MRI for the diagnosis of dentomaxillofacial malignancies in four online databases (Scopus, PubMed, Web of Science, and the Cochrane Library) was searched. Eligible for this review were original full-text articles on PET/MRI imaging, published between January 2010 and November 2024, based on experimental or clinical research involving humans. Results: Out of the 783 articles retrieved, only twelve articles were included in this systematic review. Nearly half of the articles (5 out of 12) concluded that PET/MRI is superior to PET, MRI, and PET/CT imaging in relation to defining malignancies' size. Six articles found no statistically significant results and the diagnostic accuracy presented was similar in PET/MRI versus MRI and PET/CT images. Regarding the overall risk of bias, most articles had a moderate risk. Conclusions: The use of PET/MRI in HNC cases provides a more accurate diagnosis regarding dimensions of the tumor and thus a more accurate surgical approach if needed. Further prospective studies on a larger cohort of patients are required to obtain more accurate results on the application of hybrid PET/MRI.

Accuracy of magnetic resonance imaging in the assessment of depth of invasion in tongue carcinoma: A systematic review and meta-analysis

Tongue carcinoma constitutes 10.4-46.9% of all oral squamous cell carcinomas (OSCCs) and is notoriously known for invading tissues deeper than the evident gross margins. The deeper the tumor invades, the higher are its chances of future morbidity and mortality due to extensive neck dissection and risk of recurrence. Magnetic resonance imaging (MRI) is a noninvasive diagnostic aid used for measuring a preoperative tumor's depth of invasion (DOI) as it can efficiently outline soft tissue tumors from adjacent normal tissue. To assess various MRI modalities used in measuring DOI in tongue carcinoma and their reliability compared with other DOI measuring modalities. The protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) database (CRD42022330866), and the following Preferred Reporting Items for a Systematic Review and Meta-Analysis (PRISMA) Diagnostic Test Accuracy guidelines were performed. PubMed electronic database was searched using a combination of keywords for relevant articles in the English language since 2016. Critical appraisal was carried out using the Quality Assessment of Diagnostic Accuracy Studies-Comparative (QUADAS-C) risk-of-bias (RoB) assessment tool. A weighted mean difference (WMD) was calculated between MRI and histopathological DOI along with pooled correlation and subgroup analysis, where possible. A total of 795 records were retrieved of which 17 were included in the final review with 13 included for meta-analysis. A high RoB was found for most studies for all parameters except flow and timing. WMD showed a statistically significant MRI overestimation of 1.90 mm compared with histopathology. Subgroup analysis showed the 1.5 Tesla machine to be superior to the 3.0 Tesla machine, while imaging sequence subgroup analysis could not be performed. MRI is a viable preoperative DOI measurement modality that can help in efficient treatment planning to decrease surgical morbidity and mortality.

Preoperative evaluation of depth of invasion in oral tongue squamous cell carcinoma: A systematic review and meta-analysis

The inclusion of depth of invasion (DOI) in the American Joint Committee on Cancer's staging system for oral cavity squamous cell carcinoma (SCC) has major clinical impacts. Recent studies have evaluated the reliability of imaging modalities and biopsy techniques to measure DOI preoperatively. The objective of this systematic review and meta-analysis was to comprehensively include all previously described methods to measure preoperative DOI in oral tongue SCC (OTSCC) and to compare their reliability. A systematic review was conducted on PubMed, Embase and Cochrane according to the PRISMA guidelines. Studies that evaluated the reliability of DOI measured on biopsy or imaging (rDOI) by comparing it to DOI on histopathology (pDOI) were included for extraction. A meta-analysis was conducted to obtain pooled correlation coefficients for each imaging modality. The pooled correlation coefficients between rDOI and pDOI were 0.86 (CI_95% = [0.82-0.88]) and 0.80 (CI_95% = [0.70-0.87]) for magnetic resonance imaging (MRI) studies and computed tomography (CT) studies, respectively. For ultrasound (US), the correlation coefficient could only be measured by including studies which measured not only DOI but also tumor thickness. It was 0.89 (CI_95%= [0.82-0.94]). Overall, MRI is the better studied modality. It has a good reliability to measure preoperative rDOI in OTSCC. CT is less studied but appears to be less reliable. US cannot be compared to these imaging modality as it has been used more often to measure TT than DOI.

Generative image transformer (GIT): unsupervised continuous image generative and transformable model for [123I]FP-CIT SPECT images

OBJECTIVE

Recently, generative adversarial networks began to be actively studied in the field of medical imaging. These models are used for augmenting the variation of images to improve the accuracy of computer-aided diagnosis. In this paper, we propose an alternative new image generative model based on transformer decoder blocks and verify the performance of our model in generating SPECT images that have characteristics of Parkinson's disease patients.

METHODS

Firstly, we proposed a new model architecture that is based on a transformer decoder block and is extended to generate slice images. From few superior slices of 3D volume, our model generates the rest of the inferior slices sequentially. Our model was trained by using [123I]FP-CIT SPECT images of Parkinson's disease patients that originated from the Parkinson's Progression Marker Initiative database. Pixel values of SPECT images were normalized by the specific/nonspecific binding ratio (SNBR). After training the model, we generated [123I]FP-CIT SPECT images. The transformation of images of the healthy control case SPECT images into PD-like images was also performed. Generated images were visually inspected and evaluated using the mean absolute value and asymmetric index.

RESULTS

Our model was successfully generated and transformed into PD-like SPECT images. The mean absolute SNBR was mostly less than 0.15 in absolute value. The variation of the obtained dataset images was confirmed by the analysis of the asymmetric index.

CONCLUSIONS

These results showed the potential ability of our new generative approach for SPECT images that the generative model based on the transformer realized both generation and transformation by a single model.