Iranian Journal of Medical Physics
Volume:21 Issue: 6, Nov-Dec 2024

Brain tumors (BTs) pose significant challenges in medical diagnosis and treatment due to their heterogeneity and complex characteristics. Accurate and precise segmentation of BTs in magnetic resonance images (MRIs) is crucial for effective treatment planning and patient care. In this study, we propose an ensemble deep-learning (DL) model to address the challenging task of BT segmentation. We aim to achieve accurate localization and delineation of tumor regions across different axial views. The dataset used in this study consists of 3064 T1-weighted contrast-enhanced MRI images obtained from patients diagnosed with glioma, meningioma, and pituitary tumors. Image preprocessing techniques, including normalization and intensity transformation, were applied to enhance the contrast and standardize the intensity values. The DL model is based on the DeepLabV3+ architecture combined with three well-known deep convolutional neural networks as encoders: MobileNetV2, ResNet50, and XceptionNet. The proposed ensemble model, with MobileNetV2 as the encoder, demonstrated superior performance in BT segmentation. The model achieved an average dice similarity coefficient of 0.938 and a global accuracy of 0.997. Compared to alternative models, MobileNetV2-DeepLabV3+ showed significant accuracy and segmentation precision improvements. The ensemble DL model, leveraging the strengths of MobileNetV2 and DeepLabV3+, offers a robust and efficient solution for accurate BT segmentation in MRI images. The model’s ability to delineate tumor regions holds great promise for enhancing diagnosis and treatment planning in BT analysis. Future work will explore further fine-tuning techniques and evaluate the model’s performance on larger datasets to assess its generalization capabilities.

Attenuation correction is essential for accurate PET imaging and radiotherapy (RT) dose planning. However, PET/MR systems face a significant challenge due to the lack of direct attenuation data from MR images, unlike PET/CT where CT provides inherent attenuation information. Similarly, the increasing use of MRI in RT planning necessitates pseudo-CT methods for accurate dose calculation. This study compares MR-based µ-map generation for PET/MR and pseudo-CT methods for RT planning, addressing their challenges and limitations to improve treatment accuracy and patient care.
The study involved patient selection, image processing, and generation of MR-based attenuation maps (µ-maps) for PET attenuation correction and pseudo-CTs for RT dose planning using advanced computational software.
MR-based µ-maps, potentially useful for PET attenuation correction, and pseudo CTs, potentially applicable in radiotherapy planning, were successfully generated. Head images showed MR-based µ-maps overestimating bone for two patients (deviations of 4.0% and 4.2%). Both MR-based and CT µ-maps exhibited dynamic and continuous µ-values for head bone. In the pelvis, pseudo-CT underestimated bone volume in five patients (deviations of 18.7%, 21.3%, 9.6%, 14%, and 10%). Pseudo-CT's bone µ-values lacked continuity compared to CT µ-maps. Pelvis studies revealed more dynamic and broader µ-value range for muscle in CT µ-maps than pseudo-CT and MRI µ-map.
These findings suggest the need for careful consideration and validation of attenuation correction methods, especially in regions with complex anatomical structures, to ensure accurate treatment delivery and enhance patient care in the context of PET/MR and radiotherapy.

Achieving precise dose delivery in head and neck cancer is challenging, requiring effective tumor control while minimizing toxicity. This study compares Volumetric-Modulated Arc Therapy (VMAT) and 3D Conformal Radiotherapy (3D-CRT), focusing on target coverage, Organs At Risk (OARs) sparing and treatment efficiency. For 7 randomly selected patients, Two plans (3D-CRT and VMAT) were created using Monaco TPS. The prescribed doses were 70/63/56Gy for Five patients and 69.96/59.4/54Gy for Two, in 35 and 33 fractions. The t-test was used for statistical analysis. VMAT plans underwent pretreatment quality control. The mean Dmean, Dmax, and V95% of PTV70Gy were 70.86/70.23Gy, 78.95/75.28Gy, and 95.35% / 96.97% for 3D-CRT and VMAT. The Dmax for the spinal cord, brainstem and chiasma were 49.98/40.76Gy, 64.05/50.00Gy, and 54.27/47.78Gy. The mean dose for the Left (L) and Right (R) parotids was 66.93/44.18Gy and 67.79/44.98Gy. The Dmax for L/R optic nerves and eyes were 60.17/50.41Gy, 59.44/48.49Gy, 48.4/39.88, and 44.09/36.93Gy. The 0.03cc of the L/R temporal lobes received 73.37/69.59Gy and 73.17/68.33Gy. The mean dose in 2cc of the mandible was 72.88/66.99Gy. The mean volume of the larynx with 66Gy was 23.72% / 0.76%. The Homogeneity and Conformity Index were 0.12/0.08 and 0.95/0.97. The treatment time and MUs for 3D-CRT/VMAT were 3.35/6.36 min and 806.86/621.53 MUs. VMAT gamma index passing rate was 98.6%.   The VMAT limits irradiation, reduces OARs toxicity, assures higher target dose and avoids cold and hot spots. This study shows that VMAT provides superior normal tissue protection as compared to 3D-CRT.

Bone metastasis is the advanced stage of solid malignant tumors. Bone-avid beta-emitting radiopharmaceuticals such as lutetium-177-1,4,7,10-tetraazacyclododecane-1,4,7,10-tetramethylene phosphonic acid ([177Lu]Lu-DOTMP) are effectively utilized for bone pain palliation. Radiation absorbed dose evaluation of such radiopharmaceuticals is needed in clinical works for estimating the risk associated with the usage of recently developed radiopharmaceuticals.
The radiation absorbed dose of [177Lu]Lu-DOTMP radiopharmaceutical was evaluated for adult men based on biodistribution data in Wistar rats. The Medical Internal Radiation Dosimetry (MIRD) dose calculation method and the Sparks and Aydogan methodology were applied. About 40% of the injected activity is accumulated on the surface of the trabecular and compact bones. Radiation absorbed dose of red bone marrow and osteogenic cells were estimated at 0.89±0.07 and 5.12±0.40 mGy/MBq, respectively. The maximum administrated activity was obtained at 32.2 MBq/kg (0.87 mCi/kg) of body weight with about 11.6 Gy absorbed dose of bone surface for a 70 kg adult man. The effective dose of [177Lu]Lu-DOTMP radiopharmaceutical was estimated at 0.19±0.02 mSv/MBq and the urinary bladder wall and kidneys absorbed doses were evaluated at about 0.20±0.02 mGy/MBq and 0.05±0.01 mGy/MBq, respectively.
This study indicated that [177Lu]Lu- DOTMP radiopharmaceutical can provide palliative care for bone metastases with low undesired doses to other normal tissues.

Bismuth efficiency in shielding superficial tissues in computed tomography (CT) scans has been challenged due to the imbalance between image quality and dose reduction. The aim of this study is to reduce bismuth in shields and investigate the possibility of substitution with lower atomic materials.
Five different compounds, including raw polyvinyl alcohol (PVA) and four other samples containing variable weight fractions of bismuth oxide and with a constant fraction of magnetite were selected. Shielding factors, including linear attenuation coefficient(μ), mass attenuation coefficient(μρ), half-value layer (HVL), electron density (Ne), and effective atomic number (Zeff), were evaluated over a wide range of energy by precise computational methods, XCOM, and Monte Carlo N-Particle code.
Increasing the bismuth oxide concentration improves the radiation attenuation and absorption process. This effect was observed in the μρ graphs to medium energies (E < 0.3560 MeV).The simultaneous evaluation of Ne and Zeff predicts increased absorption due to the increased and dominant photoelectric effect (<0.1000 MeV), the raised Compton effect (0.1000 < E < 0.3560 MeV; followed by scattering owing to the predominant Compton effect), and formation (E ≈ 1.2200 MeV)and dominance of pair production phenomenon (E > 3.0000 MeV).Also, a quantitative analysis of absorption through HVL in several used energies showed the efficacy of these compounds in ionizing radiation absorption.
This study establishes the advantages of substituting bismuth with compounds having lower atomic number materials, and the possibility of alleviating bismuth and replacing it with magnetite and PVA in designing radiation shields in CT scans has been confirmed.

: In order to personalize motion compensated radiotherapy with external surrogates, an intelligent method is proposed for selecting external surrogates’ motion data on the basis of patient-specific respiration pattern. This strategy enhances targeting accuracy and can potentially feed the stereoscopic X-ray imaging system and lead to fewer imaging dose, intelligently.
We investigate the effects of training data points firstly on correlation model construction at pre-treatment step for its training. Then, the same assessment will be done by means of updating data points on the model re-construction. Moreover, a recognition algorithm has been developed to detect high variability of breathing motion using pre-defined discriminator levels based on external motion amplitude.
The number of training and updating data points can be intelligently optimized depending on the breathing pattern of each patient. In addition, by developing recognition algorithm, the shooting time for motion data selection is converted from conventional strategy to intelligent approach, accordingly. As example, for a patient with high motion variability while the number of critical data points recognized by our algorithm is significant, the targeting error with and without utilizing these data points are 4.4 mm and 6.6 mm, respectively.
This work promises to be aid a more personalized delivery of motion compensated radiotherapy using external surrogates by considering to motion data gathering, according to patient-specific respiration pattern. By implementing our strategy, we expect to make a compromise between the performance accuracy of correlation model and additional imaging dose.

There is paucity of information about the impact of different neuroactivation paradigms on brain temperature changes in functional magnetic resonance spectroscopy (fMRS) studies. Magnetic resonance spectroscopy (MRS) thermometry was used to estimate the pattern of brain temperature changes with single and continuous neuroactivation paradigms.
Single-voxel MRS data was acquired from the visual cortex of four healthy volunteers using the standard spin-echo Point-RESolved Spectroscopy (PRESS) localization sequence synchronized to single and continuous visual stimulation paradigms at 3.0 tesla (T). Blood oxygenation level-dependent (BOLD) effects were estimated from changes in spectral peak height, linewidth, and area. Brain temperature was calculated by substituting the frequency offset of the water peak relative to the N-acetyl aspartate (NAA), creatine (Cr), and choline (Cho) peaks into previously deduced calibration equations for each reference peak. BOLD and temperature changes from baseline were compared by paired t-test at a significance level of p < 0.05.
In the single activation paradigm, Cho (p = 0.01) peak height, and NAA (p = 0.01) and Cr (p = 0.02) peak areas showed significant changes without significant brain temperature changes relative to all three peaks (p > 0.05). In the continuous activation paradigm, Cr (p = 0.04) peak width showed significant change, with significant brain temperature changes relative to all three reference peaks (p < 0.05).
Brain temperature significantly reduced with continuous visual activation but not with single visual activation paradigms.

The normal tissue objective (NTO), one of the new aspects in the radiation treatment planning system (TPS), aims to lower the absorbed dose received by organs at risk (OARs) close to the target volume or Planning Target Volume (PTV). This study was conducted to ascertain the impact of planning in nasopharyngeal cancer (NPC) cases both with and without manual NTO settings. The study used a 3D printed head and neck phantom exposed to radiation using the Intensity Modulated Radiation Therapy (IMRT) technique with 6000 cGy prescribed dose and divided into 30 fractions to find the discrepancies between the manually calculated absorbed dose and the automatic calculated absorbed dose of TPS. Moreover, evaluation parameter indicators, including the homogeneity index (HI), conformity index (CI), gradient index (GI), and comprehensive quality index (CQI), were used to make comparisons. NTO parameter used in manual plans are f0 = 107%, f∞ = 65%, dose fall-off  = 0.05 mm-1, and xstart = 0.75 cm. The statistical analysis resulted in a significant difference between the calculated absorbed dose and TPS's absorbed dose of Automatic NTO and Manual NTO, whereas, Without NTO plans, there was no statistical difference. The HI values for Automatic NTO, Manual NTO, and Without NTO are 0.118, 0.05 , and 0.053, respectively. The CI values for Automatic NTO, Manual NTO, and Without NTO are 0.91, 0.99, and 0,19. The GI value for Automatic NTO, Manual NTO, and Without NTO are 3.34, 4.94, and 7.29, respectively. CQI parameter showed that the Automatic NTO plan performs better than the Manual NTO plan based on the maximum dose received by the OAR. In this study, the manual NTO plan showed better performance by reducing hot spots in the central region of PTV.