Pusat Pengajian Sains Kesihatan - Tesis
Browse
Recent Submissions
- PublicationOptimising inverse planning simulated annealing (IPSA) parameters for brachytherapy planning dose calculation in oral tongue cancer patients(2025-07)Osman, Siti NoralizaBrachytherapy plays an essential role in the treatment of oral tongue cancer, providing a high radiation dose directly to the tumor while limiting the exposure to surrounding healthy tissues. Inverse Planning Simulated Annealing (IPSA) is an automated optimization algorithm commonly used in brachytherapy, but its performance can be influenced by parameter settings. This study aimed to compare the dosimetric outcomes between Graphical Optimization (GrO) and three different IPSA parameter sets in brachytherapy planning for oral tongue cancer. A retrospective analysis was conducted, where the dose-volume histogram (DVH) parameters, specifically GTV (D90) and D2cc for the mandible, pharyngeal constrictor muscle (PCM), and submandibular glands (SMG), were evaluated across different plans. The findings showed that GrO achieved the highest mean GTV D90 of 6.10 Gy, while IPSA 1, IPSA 2, and IPSA 3 recorded mean D90 values of 4.14 Gy, 4.68 Gy, and 5.37 Gy, respectively. A statistically significant differences were found in GTV D90 and D2cc mandible between GrO and IPSA plans. However, for the pharyngeal constrictor muscle, left and right submandibular glands, IPSA 2 and IPSA 3, there are no significance D2cc doses compared to GrO. Although IPSA optimization provided acceptable dose conformity, none of the IPSA parameter sets outperformed GrO in target coverage. IPSA 2 showed a favorable balance between target coverage and normal tissue sparing. As a limitation, this study only analyzed data from a single treatment fraction for each patient due to unavailable multi-fraction data, which may affect the overall interpretation. Future studies should consider collecting and analyzing data for multiple fractions to enhance accuracy and allow for more comprehensive evaluation of IPSA optimization performance in oral tongue cancer brachytherapy. Among the IPSA plans, IPSA 2 is identified as the optimal parameter configuration, offering the best overall balance between tumour coverage and OAR protection. IPSA 3, while achieving higher tumour dose coverage, resulted in increased doses to the OARs, particularly the right and left submandibular glands, highlighting the need for careful trade-off management between target coverage and normal tissue sparing in inverse planning.
- PublicationDetection and classification of breast cancer calcifications using machine learning with augmentation technique(2025-07)Borhanuddin, Nurul SyuhaidaBreast cancer calcifications are among the earliest indicators of malignancy but are often difficult to detect due to their subtle appearance and reliance on subjective radiological interpretation. This study aimed to enhance the detection and classification of breast calcifications in mammographic images through the integration of image preprocessing and augmentation techniques with machine learning. A total of 234 annotated mammographic images were collected from the Picture Archiving and Communication System at Hospital Pakar Universiti Sains Malaysia. These images were augmented using various transformations including rotation, flipping, Gaussian blur, and elastic deformation, resulting in a total dataset of 2574 images to improve variability and reduce the risk of overfitting. Preprocessing techniques such as grayscale conversion, contrast enhancement using CLAHE, and top hat filtering were applied to improve the visibility of calcification features. Five machine learning models were evaluated including Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Random Forest (RF), Logistic Regression (LR), and a Soft Voting ensemble model. Model performance was measured using accuracy, precision, recall, specificity, and F1 score. Validation was performed using 5-fold cross validation and statistical significance was tested with the Friedman test and Wilcoxon Signed Rank test. Based on the results, the KNN model achieved the highest average accuracy of 87.61% followed by SVM at 79.07%, RF at 78.64% and LR at 69.62%. The findings suggest that the KNN model was particularly effective at distinguishing between benign and malignant calcifications due to its sensitivity to local feature patterns. Although Logistic Regression had the shortest training time, it performed the poorest in all evaluation metrics, indicating that training speed alone is not a sufficient measure of diagnostic utility. The results also highlight that proper augmentation and preprocessing not only improve model accuracy but also contribute to more balanced performance across sensitivity and specificity. The use of statistical validation confirmed that differences among the model performances were significant, thus reinforcing the reliability of the findings. This study demonstrates that machine learning models when supported by proper data preparation strategies, can serve as effective tools in the development of computer assisted diagnostic systems for early breast cancer detection
- PublicationInvestigate the effect of ortopedic metal artefact reduction algorithm on dosimetry of imrt prostate with hip prosthesis(2025-07)Zakir, Nurul Najah MohdProstate cancer is one of the most common cancers among Malaysian men, and radiotherapy—especially intensity-modulated radiation therapy (IMRT)—is a key treatment. However, metal hip implants in some patients can cause artefacts in computed tomography (CT) images, which affect the accuracy of anatomy and radiation dose calculations. This study investigated the efficiency of the Orthopedic Metal Artifact Reduction (OMAR) algorithm by Philips reduces these artefacts and improves dose accuracy in prostate IMRT. A custom male pelvic phantom with a metal hip implant was created using tissue-like materials and treated with a 6 MV photon beam. The radiation doses were calculated using the Eclipse treatment planning system (TPS) and checked using a calibrated thermoluminescent dosimeter (TLD) type 100. The results showed that OMAR significantly reduced image distortion and improved the accuracy of structure outlines and dose delivery. Treatment plans based on OMAR-corrected images had better dose conformity and uniformity. Statistical tests confirmed significant improvements in dose accuracy (p > 0.05) in several organs. Hence, the OMAR algorithm has been proven that it improve both image quality and radiation dose accuracy, making it a valuable tool for radiotherapy planning in patients with metal implants.
- PublicationEvaluation of varian ix clinical linear accelerator performance based on daily quality assurance using quickcheck phantom(2025-07)Yusniman, Nurul AfiqahThis study aims to evaluate the performance consistency of the Varian iX Clinical Linear Accelerator (LINAC) by analysing radiation output parameters using the PTW QUICKCHECK Phantom during daily quality assurance (QA) procedures at Hospital Pakar Universiti Sains Malaysia (HPUSM). The parameters measured include central axis dose (CAX), beam flatness, symmetry in the gantry-target (GT) and left-right (LR) directions, and beam quality factor (BQF) for photon (6 MV) and electron beams (6, 9, 12, and 15 MeV). Data were collected over a three-month period from January to March 2025. All measurements were taken using a 10 × 10 cm² field size with 100 MU delivered at a dose rate of 300 MU/min. Repeated measurements were conducted in cases of reading errors, and any invalid data were excluded from the analysis. The results showed that all radiation output parameters remained within the tolerance limits recommended by AAPM TG-142 (±3%), indicating excellent output stability and machine performance. This confirms the suitability of the QUICKCHECK device as a reliable and efficient tool for routine daily QA in radiotherapy settings.
- PublicationInvestigation of free breathing technique implication towards cardiac structure dose in left breast radiotherapy: retrospective study(2025-07)Shahari, Nur SyahminaThis study aimed to comprehensively evaluate the dosimetric parameters of cardiac structures, assess volumetric-dosimetric correlations, and investigate the relationship between heart dose and LAD specific dosimetry in left breast radiotherapy plans. The objective was to ascertain adherence to recommended tolerance levels and identify critical parameters for cardiac sparing. A retrospective analysis was conducted on 35 left breast cancer patients treated with 3D-CRT at Hospital Pakar Universiti Sains Malaysia (HPUSM) from January to December 2024, receiving a prescribed dose of 40 Gy in 15 fractions. No data beyond December 2024 was included in this study, and the research period did not extend beyond July 2025. Ethical clearance was secured, and purposive sampling was employed. CT simulation was performed under free breathing conditions. Organs at risk (OARs), including the heart, underwent auto-contouring, while the LAD was manually delineated by a medical physicist due to its complicated structure visualized in the CT images. Dosimetric parameters were assessed against RTOG 0617 and DEGRO tolerance guidelines. Spearman's rho correlation analysis was utilized to determine relationships between heart/LAD volumes and their respective doses, as well as between heart mean dose and LAD-specific doses, following preliminary normality testing. The analysis revealed no statistically significant correlation between heart volume and heart mean dose (r =0.018, p=0.917), nor between LAD volume and its dosimetric parameters (LAD Dmean: r =−0.077, p=0.659; LAD V40: r =0.162, p=0.353). A moderate, statistically significant positive correlation was identified between heart mean dose and mean LAD dose (r =0.442, p=0.008). Conversely, the correlation between heart mean dose and LAD V40 (%) was not statistically significant (r =0.240, p=0.165). Volumetric parameters of the heart and LAD do not reliably predict radiation dose to these structures. While a correlation exists between heart mean dose and mean LAD dose, global heart dose metrics are insufficient to fully capture critical high-dose regions within the LAD. Consequently, for effective cardiac sparing in left breast radiotherapy, future planning necessitates meticulous monitoring of both the overall heart mean dose and, crucially, LAD-specific parameters, particularly LAD V40 (%), to ensure comprehensive cardiac protection and mitigate long-term cardiotoxicity. Future investigations should involve larger cohorts to confirm these findings and explore the clinical impact of LAD V40 exceedance.