A Deep Learning Algorithm for Radiographic Measurements of the Hip in Adults-A Reliability and Agreement Study

Hip dysplasia (HD) is a frequent cause of hip pain in skeletally mature patients and may lead to osteoarthritis (OA). An accurate and early diagnosis may postpone, reduce or even prevent the onset of OA and ultimately hip arthroplasty at a young age. The overall aim of this study was to assess the reliability of an algorithm, designed to read pelvic anterior-posterior (AP) radiographs and to estimate the agreement between the algorithm and human readers for measuring (i) lateral center edge angle of Wiberg (LCEA) and (ii) Acetabular index angle (AIA). The algorithm was based on deep-learning models developed using a modified U-net architecture and ResNet 34. The newly developed algorithm was found to be highly reliable when identifying the anatomical landmarks used for measuring LCEA and AIA in pelvic radiographs, thus offering highly consistent measurement outputs. The study showed that manual identification of the same landmarks made by five specialist readers were subject to variance and the level of agreement between the algorithm and human readers was consequently poor with mean measured differences from 0.37 to 9.56° for right LCEA measurements. The algorithm displayed the highest agreement with the senior orthopedic surgeon. With further development, the algorithm may be a good alternative to humans when screening for HD.

Diagnostic Accuracy and Clinical Impact of [ 18F]FET PET in Childhood CNS tumors

BACKGROUND

Central nervous system (CNS) tumors cause the highest death rates among childhood cancers, and survivors frequently have severe late effects. Magnetic resonance imaging (MRI) is the imaging modality of choice, but its specificity can be challenged by treatment-induced signal changes. In adults, O-(2-[ 18F]fluoroethyl)-L-tyrosine ([ 18F]FET) PET can assist in interpreting MRI findings. We assessed the clinical impact and diagnostic accuracy of adding [ 18F]FET PET to MRI in children with CNS tumors.

METHODS

A total of 169 [ 18F]FET PET scans were performed in 97 prospectively and consecutively included patients with known or suspected childhood CNS tumors. Scans were performed at primary diagnosis, before or after treatment, or at relapse.

RESULTS

Adding [ 18F]FET PET to MRI impacted clinical management in 8% [95% confidence interval (CI): 4-13%] of all scans (n=151) and in 33% [CI: 17-53%] of scans deemed clinically indicated due to difficult decision-making on MRI alone (n=30). Using pathology or follow-up as reference standard, the addition of [ 18F]FET PET increased specificity (1.00 [0.82-1.00] vs. 0.48 [0.30-0.70], p=0.0001) and accuracy (0.91 [CI: 0.87-0.96] vs. 0.81 [CI: 0.75-0.89], p=0.04) in 83 treated lesions and accuracy in 58 untreated lesions (0.96 [CI:0.91-1.00] vs 0.90 [CI:0.82-0.92], p<0.001). Further, in a subset of patients (n=15) [ 18F]FET uptake correlated positively with genomic proliferation index.

CONCLUSIONS

The addition of [ 18F]FET PET to MRI helped discriminate tumor from non-tumor lesions in the largest consecutive cohort of pediatric CNS tumor patients presented to date.

Early Postoperative 18F-FET PET/MRI for Pediatric Brain and Spinal Cord Tumors

Complete resection is the treatment of choice for most pediatric brain tumors, but early postoperative MRI for detection of residual tumor may be misleading because of MRI signal changes caused by the operation. PET imaging with amino acid tracers in adults increases the diagnostic accuracy for brain tumors, but the literature in pediatric neurooncology is limited. A hybrid PET/MRI system is highly beneficial in children, reducing the number of scanning procedures, and this is to our knowledge the first larger study using PET/MRI in pediatric neurooncology. We evaluated if additional postoperative ¹⁸F-fluoro-ethyl-tyrosine (¹⁸F-FET) PET in children and adolescents would improve diagnostic accuracy for the detection of residual tumor as compared with MRI alone and would assist clinical management. Methods: Twenty-two patients (7 male; mean age, 9.5 y; range, 0-19 y) were included prospectively and consecutively in the study and had 27 early postoperative ¹⁸F-FET PET exams performed preferentially in a hybrid PET/MRI system (NCT03402425). Results: Using follow-up (93%) or reoperation (7%) as the reference standard, PET combined with MRI discriminated tumor from treatment effects with a lesion-based sensitivity/specificity/accuracy (95% confidence intervals) of 0.73 (0.50-1.00)/1.00 (0.74-1.00)/0.87 (0.73-1.00) compared with MRI alone: 0.80 (0.57-1.00)/0.75 (0.53-0.94)/0.77 (0.65-0.90); that is, the specificity for PET/MRI was 1.00 as compared with 0.75 for MRI alone (P = 0.13). In 11 of 27 cases (41%), results from the ¹⁸F-FET PET scans added relevant clinical information, including one scan that directly influenced clinical management because an additional residual tumor site was identified. ¹⁸F-FET uptake in reactive changes was frequent (52%), but correct interpretation was possible in all cases. Conclusion: The high specificity for detecting residual tumor suggests that supplementary ¹⁸F-FET PET is relevant in cases where reoperation for residual tumor is considered.

Feasibility of multi-parametric PET and MRI for prediction of tumour recurrence in patients with glioblastoma

BACKGROUND

Recurrence in glioblastoma patients often occur close to the original tumour and indicates that the current treatment is inadequate for local tumour control. In this study, we explored the feasibility of using multi-modality imaging at the time of radiotherapy planning. Specifically, we aimed to identify parameters from pre-treatment PET and MRI with potential to predict tumour recurrence.

MATERIALS AND METHODS

Sixteen patients were prospectively recruited and treated according to established guidelines. Multi-parametric imaging with ¹⁸F-FET PET/CT and ¹⁸F-FDG PET/MR including diffusion and dynamic contrast enhanced perfusion MRI were performed before radiotherapy. Correlations between imaging parameters were calculated. Imaging was related to the voxel-wise outcome at the time of tumour recurrence. Within the radiotherapy target, median differences of imaging parameters in recurring and non-recurring voxels were calculated for contrast-enhancing lesion (CEL), non-enhancing lesion (NEL), and normal appearing grey and white matter. Logistic regression models were created to predict the patient-specific probability of recurrence. The most important parameters were identified using standardized model coefficients.

RESULTS

Significant median differences between recurring and non-recurring voxels were observed for FDG, FET, fractional anisotropy, mean diffusivity, mean transit time, extra-vascular, extra-cellular blood volume and permeability derived from scans prior to chemo-radiotherapy. Tissue-specific patterns of voxel-wise correlations were observed. The most pronounced correlations were observed for ¹⁸F-FDG- and ¹⁸F-FET-uptake in CEL and NEL. Voxel-wise modelling of recurrence probability resulted in area under the receiver operating characteristic curve of 0.77 from scans prior to therapy. Overall, FET proved to be the most important parameter for recurrence prediction.

CONCLUSION

Multi-parametric imaging before radiotherapy is feasible and significant differences in imaging parameters between recurring and non-recurring voxels were observed. Combining parameters in a logistic regression model enabled patient-specific maps of recurrence probability, where ¹⁸F-FET proved to be most important. This strategy could enable risk-adapted radiotherapy planning.

Patterns of failure for patients with glioblastoma following O-(2-[18F]fluoroethyl)-L-tyrosine PET- and MRI-guided radiotherapy

BACKGROUND AND PURPOSE

To evaluate the patterns of failure following clinical introduction of amino-acid O-(2-[¹⁸F]fluoroethyl)-L-tyrosine (FET)-PET-guided target definition for radiotherapy (RT) of glioblastoma patients.

MATERIALS AND METHODS

The first 66 consecutive patients with confirmed histology, scanned using FET-PET/CT and MRI were selected for evaluation. Chemo-radiotherapy was delivered to a volume based on both MRI and FET-PET (PET_vol). The volume of recurrence (RV) was defined on MRI data collected at the time of progression according to RANO criteria.

RESULTS

Fifty patients were evaluable, with median follow-up of 45months. Central, in-field, marginal and distant recurrences were observed for 82%, 10%, 2%, and 6% of the patients, respectively. We found a volumetric overlap of 26%, 31% and 39% of the RV with the contrast-enhancing MR volume, PET_vol and the composite MRPET_vol, respectively. MGMT-methylation (p=0.03), larger PET_vol (p<0.001), and less extensive surgery (p<0.001), were associated with larger PET_vol overlap.

CONCLUSION

The combined MRPET_vol had a stronger association with the recurrence volume than either of the modalities alone. Larger overlap of PET_vol and RV was observed for patients with MGMT-methylation, less extensive surgery, and large PET_vol on the RT-planning scans.

Clinical PET/MRI in neurooncology: opportunities and challenges from a single-institution perspective

Purpose

Magnetic resonance imaging (MRI) plays a key role in neurooncology, i.e., for diagnosis, treatment evaluation and detection of recurrence. However, standard MRI cannot always separate malignant tissue from other pathologies or treatment-induced changes. Advanced MRI techniques such as diffusion-weighted imaging, perfusion imaging and spectroscopy show promising results in discriminating malignant from benign lesions. Further, supplemental imaging with amino acid positron emission tomography (PET) has been shown to increase accuracy significantly and is used routinely at an increasing number of sites. Several centers are now implementing hybrid PET/MRI systems allowing for multiparametric imaging, combining conventional MRI with advanced MRI and amino acid PET imaging. Neurooncology is an obvious focus area for PET/MR imaging.

Methods

Based on the literature and our experience from more than 300 PET/MRI examinations of brain tumors with ¹⁸F-fluoro-ethyl-tyrosine, the clinical use of PET/MRI in adult and pediatric neurooncology is critically reviewed.

Results

Although the results are increasingly promising, the added value and range of indications for multiparametric imaging with PET/MRI are yet to be established. Robust solutions to overcome the number of issues when using a PET/MRI scanner are being developed, which is promising for a more routine use in the future.

Conclusions

In a clinical setting, a PET/MRI scan may increase accuracy in discriminating recurrence from treatment changes, although sequential same-day imaging on separate systems will often constitute a reliable and cost-effective alternative. Pediatric patients who require general anesthesia will benefit the most from simultaneous PET and MR imaging.

Early changes in perfusion of glioblastoma during radio- and chemotherapy evaluated by T1-dynamic contrast enhanced magnetic resonance imaging

BACKGROUND

The survival times of patients with glioblastoma differ widely and biomarkers that would enable individualized treatment are needed. The objective of this study was to measure changes in the vascular physiology of tumor using T1-dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) in patients with glioblastoma during early stages of radio- and chemotherapy (Tx) and explore possible correlations with treatment outcomes.

MATERIAL AND METHODS

An exploratory prospective study was planned. Patients underwent DCE-MRI at baseline, after approximately one and six weeks of Tx and three and six months post-Tx. DCE-MRI at three Tesla generated maps of blood flow (BF), blood volume (BV), permeability (Ki) and volume of distribution (Vd) using a combination of model-free deconvolution and Patlak plots. Regions of interest in contrast enhancing tumor and in normal appearing white matter were contoured. Progression-free survival (PFS) was the primary clinical outcome. Patients with PFS > 6 months were compared with those with PFS < 6 months. Parameters of vascular physiology and changes in these during Tx were compared for these two groups at all time points using non-parametric statistics.

RESULTS

Eleven eligible patients were included and 46 DCE-MRI examinations were carried out. BF in tumor increased for all patients early during Tx (p = 0.005) and then fell to a level below baseline at post-Tx examinations (p = 0.016). A similar but non-significant trend was seen for tumor BV. There was no detectable difference between patients with PFS > 6 months versus PFS < 6 months with regards to baseline values or changes during and after Tx.

CONCLUSIONS

Although no correlations to outcomes were found, the results of this exploratory study may be hypothesis generating and will be examined in a larger patient group.

Dose painting based on tumor uptake of Cu-ATSM and FDG: a comparative study

BACKGROUND

Hypoxia and increased glycolytic activity of tumors are associated with poor prognosis. The purpose of this study was to investigate differences in radiotherapy (RT) dose painting based on the uptake of 2-deoxy-2-[(18) F]-fluorodeoxyglucose (FDG) and the proposed hypoxia tracer, copper(II)diacetyl-bis(N(4))-methylsemithiocarbazone (Cu-ATSM) using spontaneous clinical canine tumor models.

METHODS

Positron emission tomography/computed tomography scans of five spontaneous canine sarcomas and carcinomas were obtained; FDG on day 1 and (64)Cu-ATSM on day 2 and 3 (approx. 3 and 24 hours pi.). Sub-volumes for dose escalation were defined by a threshold-based method for both tracers and five dose escalation levels were formed in each sub-volume. Volumetric modulated arc therapy plans were optimized based on the dose escalation regions for each scan for a total of three dose plans for each dog. The prescription dose for the GTV was 45 Gy (100%) and it was linearly escalated to a maximum of 150%. The correlations between dose painting plans were analyzed with construction of dose distribution density maps and quality volume histograms (QVH). Correlation between high-dose regions was investigated with Dice correlation coefficients.

RESULTS

Comparison of dose plans revealed varying degree of correlation between cases. Some cases displayed a separation of high-dose regions in the comparison of FDG vs. (64)Cu-ATSM dose plans at both time points. Among the Dice correlation coefficients, the high dose regions showed the lowest degree of agreement, indicating potential benefit of using multiple tracers for dose painting. QVH analysis revealed that FDG-based dose painting plans adequately covered approximately 50% of the hypoxic regions.

CONCLUSION

Radiotherapy plans optimized with the current approach for cut-off values and dose region definitions based on FDG, (64)Cu-ATSM 3 h and 24 h uptake in canine tumors had different localization of the regional dose escalation levels. This indicates that (64)Cu-ATSM at two different time-points and FDG provide different biological information that has to be taken into account when using the dose painting strategy in radiotherapy treatment planning.