The Effect of Image Resampling on the Performance of Radiomics-Based Artificial Intelligence in Multicenter Prostate MRI

BACKGROUND

Single center MRI radiomics models are sensitive to data heterogeneity, limiting the diagnostic capabilities of current prostate cancer (PCa) radiomics models.

PURPOSE

To study the impact of image resampling on the diagnostic performance of radiomics in a multicenter prostate MRI setting.

STUDY TYPE

Retrospective.

POPULATION

Nine hundred thirty patients (nine centers, two vendors) with 737 eligible PCa lesions, randomly split into training (70%, N = 500), validation (10%, N = 89), and a held-out test set (20%, N = 148).

FIELD STRENGTH/SEQUENCE

1.5T and 3T scanners/T2-weighted imaging (T2W), diffusion-weighted imaging (DWI), and apparent diffusion coefficient maps.

ASSESSMENT

A total of 48 normalized radiomics datasets were created using various resampling methods, including different target resolutions (T2W: 0.35, 0.5, and 0.8 mm; DWI: 1.37, 2, and 2.5 mm), dimensionalities (2D/3D) and interpolation techniques (nearest neighbor, linear, Bspline and Blackman windowed-sinc). Each of the datasets was used to train a radiomics model to detect clinically relevant PCa (International Society of Urological Pathology grade ≥ 2). Baseline models were constructed using 2D and 3D datasets without image resampling. The resampling configurations with highest validation performance were evaluated in the test dataset and compared to the baseline models.

STATISTICAL TESTS

Area under the curve (AUC), DeLong test. The significance level used was 0.05.

RESULTS

The best 2D resampling model (T2W: Bspline and 0.5 mm resolution, DWI: nearest neighbor and 2 mm resolution) significantly outperformed the 2D baseline (AUC: 0.77 vs. 0.64). The best 3D resampling model (T2W: linear and 0.8 mm resolution, DWI: nearest neighbor and 2.5 mm resolution) significantly outperformed the 3D baseline (AUC: 0.79 vs. 0.67).

DATA CONCLUSION

Image resampling has a significant effect on the performance of multicenter radiomics artificial intelligence in prostate MRI. The recommended 2D resampling configuration is isotropic resampling with T2W at 0.5 mm (Bspline interpolation) and DWI at 2 mm (nearest neighbor interpolation). For the 3D radiomics, this work recommends isotropic resampling with T2W at 0.8 mm (linear interpolation) and DWI at 2.5 mm (nearest neighbor interpolation).

EVIDENCE LEVEL

3 TECHNICAL EFFICACY: Stage 2.

Using deep learning to optimize the prostate MRI protocol by assessing the diagnostic efficacy of MRI sequences

PURPOSE

To explore diagnostic deep learning for optimizing the prostate MRI protocol by assessing the diagnostic efficacy of MRI sequences.

METHOD

This retrospective study included 840 patients with a biparametric prostate MRI scan. The MRI protocol included a T2-weighted image, three DWI sequences (b50, b400, and b800 s/mm2), a calculated ADC map, and a calculated b1400 sequence. Two accelerated MRI protocols were simulated, using only two acquired b-values to calculate the ADC and b1400. Deep learning models were trained to detect prostate cancer lesions on accelerated and full protocols. The diagnostic performances of the protocols were compared on the patient-level with the area under the receiver operating characteristic (AUROC), using DeLong's test, and on the lesion-level with the partial area under the free response operating characteristic (pAUFROC), using a permutation test. Validation of the results was performed among expert radiologists.

RESULTS

No significant differences in diagnostic performance were found between the accelerated protocols and the full bpMRI baseline. Omitting b800 reduced 53% DWI scan time, with a performance difference of + 0.01 AUROC (p = 0.20) and -0.03 pAUFROC (p = 0.45). Omitting b400 reduced 32% DWI scan time, with a performance difference of -0.01 AUROC (p = 0.65) and + 0.01 pAUFROC (p = 0.73). Multiple expert radiologists underlined the findings.

CONCLUSIONS

This study shows that deep learning can assess the diagnostic efficacy of MRI sequences by comparing prostate MRI protocols on diagnostic accuracy. Omitting either the b400 or the b800 DWI sequence can optimize the prostate MRI protocol by reducing scan time without compromising diagnostic quality.

Does FDG-PET/CT for incidentally found pulmonary lesions lead to a cascade of more incidental findings?

OBJECTIVE

To determine the frequency, nature, and downstream healthcare costs of new incidental findings that are found on whole-body FDG-PET/CT in patients with a non-FDG-avid pulmonary lesion ≥10 mm that was incidentally found on previous imaging.

MATERIALS AND METHODS

This retrospective study included a consecutive series of patients who underwent whole-body FDG-PET/CT because of an incidentally found pulmonary lesion ≥10 mm.

RESULTS

Seventy patients were included, of whom 23 (32.9 %) had an incidentally found pulmonary lesion that proved to be non-FDG-avid. In 12 of these 23 cases (52.2 %) at least one new incidental finding was discovered on FDG-PET/CT. The total number of new incidental findings was 21, of which 7 turned out to be benign, 1 proved to be malignant (incurable metastasized cancer), and 13 whose nature remained unclear. One patient sustained permanent neurologic impairment of the left leg due to iatrogenic nerve damage during laparotomy for an incidental finding which turned out to be benign. The total costs of all additional investigations due to the detection of new incidental findings amounted to €9903.17, translating to an average of €141.47 per whole-body FDG-PET/CT scan performed for the evaluation of an incidentally found pulmonary lesion.

CONCLUSION

In many patients in whom whole-body FDG-PET/CT was performed to evaluate an incidentally found pulmonary lesion that turned out to be non-FDG-avid and therefore very likely benign, FDG-PET/CT detected new incidental findings in our preliminary study. Whether the detection of these new incidental findings is cost-effective or not, requires further research with larger sample sizes.

Assessing Authorship Rates over Time in Original Radiologic Research Publications

Background Previous studies have shown an increase in the number of authors on radiologic articles between 1950 and 2013, but the cause is unclear. Purpose To determine whether authorship rate in radiologic and general medical literature has continued to increase and to assess study variables associated with increased author numbers. Materials and Methods PubMed/Medline was searched for articles published between January 1998 and October 2022 in general radiology and general medical journals with the top five highest current impact factors. Generalized linear regression analysis was used to calculate adjusted incidence rate ratios (IRRs) for the numbers of authors. Wald tests assessed the associations between study variables and the numbers of authors per article. Combined mixed-effects regression analysis was performed to compare general medicine and radiology journals. Results There were 3381 original radiologic research articles that were analyzed. Authorship rate increased between 1998 (median, six authors; IQR, 4) and 2022 (median, 11 authors; IQR, 8). Later publication year was associated with more authors per article (IRR, 1.02; 95% CI: 1.01, 1.02; P < .001) after adjusting for publishing journal, continent of origin of first author, number of countries involved, PubMed/Medline original article type, study design, number of disciplines involved, multicenter or single-center study, reporting of a priori power calculation, reporting of obtaining informed consent, study sample size, and number of article pages. There were 1250 general medicine original research articles that were analyzed. Later publication year was also associated with more authors after adjustment for the study variables (IRR, 1.04; 95% CI: 1.03, 1.05; P < .001). There was a stronger increase in authorship by publication year for general medicine journals compared with radiology journals (IRR, 1.02; 95% CI: 1.01, 1.02; P < .001). Conclusion An increase in authorship rate was observed in the radiologic and general medical literature between 1998 and 2022, and the number of authors per article was independently associated with later year of publication. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Arrivé in this issue.

Focused view CT angiography for selective visualization of stroke related arteries: technical feasibility

OBJECTIVES

This study investigated the technical feasibility of focused view CTA for the selective visualization of stroke related arteries.

METHODS

A total of 141 CTA examinations for acute ischemic stroke evaluation were divided into a set of 100 cases to train a deep learning algorithm (dubbed "focused view CTA") that selectively extracts brain (including intracranial arteries) and extracranial arteries, and a test set of 41 cases. The visibility of anatomic structures at focused view and unmodified CTA was assessed using the following scoring system: 5 = completely visible, diagnostically sufficient; 4 = nearly completely visible, diagnostically sufficient; 3 = incompletely visible, barely diagnostically sufficient; 2 = hardly visible, diagnostically insufficient; 1 = not visible, diagnostically insufficient.

RESULTS

At focused view CTA, median scores for the aortic arch, subclavian arteries, common carotid arteries, C1, C6, and C7 segments of the internal carotid arteries, V4 segment of the vertebral arteries, basilar artery, cerebellum including cerebellar arteries, cerebrum including cerebral arteries, and dural venous sinuses, were all 4. Median scores for the C2 to C5 segments of the internal carotid arteries, and V1 to V3 segments of the vertebral arteries ranged between 3 and 2. At unmodified CTA, median score for all above-mentioned anatomic structures was 5, which was significantly higher (p < 0.0001) than that at focused view CTA.

CONCLUSION

Focused view CTA shows promise for the selective visualization of stroke-related arteries. Further improvements should focus on more accurately visualizing the smaller and tortuous internal carotid and vertebral artery segments close to bone.

CLINICAL RELEVANCE

Focused view CTA may speed up image interpretation time for LVO detection and may potentially be used as a tool to study the clinical relevance of incidental findings in future prospective long-term follow-up studies.

KEY POINTS

• A deep learning-based algorithm ("focused view CTA") was developed to selectively visualize relevant structures for acute ischemic stroke evaluation at CTA. • The elimination of unrequested anatomic background information was complete in all cases. • Focused view CTA may be used to study the clinical relevance of incidental findings.

Can we revolutionize diagnostic imaging by keeping Pandora's box closed?

Incidental imaging findings are a considerable health problem, because they generally result in low-value and potentially harmful care. Healthcare professionals struggle how to deal with them, because once detected they can usually not be ignored. In this opinion article, we first reflect on current practice, and then propose and discuss a new potential strategy to pre-emptively tackle incidental findings. The core principle of this concept is to keep the proverbial Pandora's box closed, i.e. to not visualize incidental findings, which can be achieved using deep learning algorithms. This concept may have profound implications for diagnostic imaging.

A new medical imaging postprocessing and interpretation concept to investigate the clinical relevance of incidentalomas: can we keep Pandora's box closed?

BACKGROUND

Incidental imaging findings (incidentalomas) are common, but there is currently no effective means to investigate their clinical relevance.

PURPOSE

To introduce a new concept to postprocess a medical imaging examination in a way that incidentalomas are concealed while its diagnostic potential is maintained to answer the referring physician's clinical questions.

MATERIAL AND METHODS

A deep learning algorithm was developed to automatically eliminate liver, gallbladder, pancreas, spleen, adrenal glands, lungs, and bone from unenhanced computed tomography (CT). This deep learning algorithm was applied to a separately held set of unenhanced CT scans of 27 patients who underwent CT to evaluate for urolithiasis, and who had a total of 32 incidentalomas in one of the aforementioned organs.

RESULTS

Median visual scores for organ elimination on modified CT were 100% for the liver, gallbladder, spleen, and right adrenal gland, 90%-99% for the pancreas, lungs, and bones, and 80%-89% for the left adrenal gland. In 26 out of 27 cases (96.3%), the renal calyces and pelves, ureters, and urinary bladder were completely visible on modified CT. In one case, a short (<1 cm) trajectory of the left ureter was not clearly visible due to adjacent atherosclerosis that was mistaken for bone by the algorithm. Of 32 incidentalomas, 28 (87.5%) were completely concealed on modified CT.

CONCLUSION

This preliminary technical report demonstrated the feasibility of a new approach to postprocess and evaluate medical imaging examinations that can be used by future prospective research studies with long-term follow-up to investigate the clinical relevance of incidentalomas.

Comparative Performance of Deep Learning and Radiologists for the Diagnosis and Localization of Clinically Significant Prostate Cancer at MRI: A Systematic Review

Background: Deep learning (DL)-based models have demonstrated an ability to automatically diagnose clinically significant prostate cancer (PCa) on MRI scans and are regularly reported to approach expert performance. The aim of this work was to systematically review the literature comparing deep learning (DL) systems to radiologists in order to evaluate the comparative performance of current state-of-the-art deep learning models and radiologists. Methods: This systematic review was conducted in accordance with the 2020 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist. Studies investigating DL models for diagnosing clinically significant (cs) PCa on MRI were included. The quality and risk of bias of each study were assessed using the checklist for AI in medical imaging (CLAIM) and QUADAS-2, respectively. Patient level and lesion-based diagnostic performance were separately evaluated by comparing the sensitivity achieved by DL and radiologists at an identical specificity and the false positives per patient, respectively. Results: The final selection consisted of eight studies with a combined 7337 patients. The median study quality with CLAIM was 74.1% (IQR: 70.6–77.6). DL achieved an identical patient-level performance to the radiologists for PI-RADS ≥ 3 (both 97.7%, SD = 2.1%). DL had a lower sensitivity for PI-RADS ≥ 4 (84.2% vs. 88.8%, p = 0.43). The sensitivity of DL for lesion localization was also between 2% and 12.5% lower than that of the radiologists. Conclusions: DL models for the diagnosis of csPCa on MRI appear to approach the performance of experts but currently have a lower sensitivity compared to experienced radiologists. There is a need for studies with larger datasets and for validation on external data.

AI-assisted biparametric MRI surveillance of prostate cancer: feasibility study

OBJECTIVES

To evaluate the feasibility of automatic longitudinal analysis of consecutive biparametric MRI (bpMRI) scans to detect clinically significant (cs) prostate cancer (PCa).

METHODS

This retrospective study included a multi-center dataset of 1513 patients who underwent bpMRI (T2 + DWI) between 2014 and 2020, of whom 73 patients underwent at least two consecutive bpMRI scans and repeat biopsies. A deep learning PCa detection model was developed to produce a heatmap of all PIRADS ≥ 2 lesions across prior and current studies. The heatmaps for each patient's prior and current examination were used to extract differential volumetric and likelihood features reflecting explainable changes between examinations. A machine learning classifier was trained to predict from these features csPCa (ISUP > 1) at the current examination according to biopsy. A classifier trained on the current study only was developed for comparison. An extended classifier was developed to incorporate clinical parameters (PSA, PSA density, and age). The cross-validated diagnostic accuracies were compared using ROC analysis. The diagnostic performance of the best model was compared to the radiologist scores.

RESULTS

The model including prior and current study (AUC 0.81, CI: 0.69, 0.91) resulted in a higher (p = 0.04) diagnostic accuracy than the current only model (AUC 0.73, CI: 0.61, 0.84). Adding clinical variables further improved diagnostic performance (AUC 0.86, CI: 0.77, 0.93). The diagnostic performance of the surveillance AI model was significantly better (p = 0.02) than of radiologists (AUC 0.69, CI: 0.54, 0.81).

CONCLUSIONS

Our proposed AI-assisted surveillance of prostate MRI can pick up explainable, diagnostically relevant changes with promising diagnostic accuracy.

KEY POINTS

• Sequential prostate MRI scans can be automatically evaluated using a hybrid deep learning and machine learning approach. • The diagnostic accuracy of our csPCa detection AI model improved by including clinical parameters.

Daily rat tibial growth in vivo following hypothalamic sex reversal with neonatal and pubertal treatments with gonadal steroids

A striking sex-related difference in postpubertal growth and growth hormone (GH) secretory pattern in the rat has been described. Although this sexual dimorphism seems to be determined by the neonatal effects of gonadal steroids on the hypothalamus, peripubertal exposure to steroids also plays an important role. In order to study the real influence of the hypothalamic sex and/or peripubertal gonadal steroids, the growth pattern of female and male rats in response to neonatal and peripubertal sexual steroid treatments was studied using microknemometry, a technique that allows non-invasive daily measurements of rat tibial growth rate. Neonatal steroid environment in males was modified by castration on day 1, whereas in females it was changed by a single neonatal testosterone administration on day 5 followed by castration at 13 days of age. From the onset of puberty to adulthood, both female and male animals received testosterone or estrogens, respectively. Neonatal treatment alone, i.e. androgenization of female and castration of male rats, were only able to induce a partial reversal of the original sex-dependent growth pattern. Additional peripubertal treatments achieved a complete change in the sex-linked growth pattern. Consistent with the effects observed on growth, the pituitary GH concentration was significantly increased in females, and diminished in males, when they were treated both at the neonatal and peripubertal stages. However, only this latter group, whose growth was more seriously compromised, showed decreased plasma insulin-like growth factor-I (IGF-I) levels. In conclusion, a complete feminization of male tibial growth pattern or masculinization of female pattern can only be achieved by maintaining the new steroid environment from puberty to adulthood.