Hasil Pencarian  ::  Simpan CSV :: Kembali

"Latar Belakang: Insiden kanker sel skuamosa rongga mulut (KSSRM) tercatat meningkat pada dekade terakhir, sebagian besar kasus datang dengan stadium lokal lanjut. Kemoterapi induksi merupakan rekomendasi pada sebagian besar kasus stadium lokal lanjut, dengan harapan tumor mengecil sehingga dapat dioperasi lebih baik dan dapat dilakukan preservasi organ. Namun tercatat, 40 % pasien tidak berespon baik terhadap kemoterapi induksi, sehingga berpeluang menambah morbiditas bahkan mengubah status tumor menjadi unresectable. Nilai Apparent Diffusion Coefficient (ADC) pada Diffusion Weighted - MRI (DW-MRI) merupakan parameter fungsional MRI yang berhubungan dengan densitas sel jaringan kondisi matriks ekstraseluler. Sehingga pemeriksaan ini dapat digunakan untuk memprediksi respon kemoterapi pasien.Tujuan: mengetahui hubungan nilai ADC pra pengobatan dengan respon kemoterapi pada pasien KSSRM stadium lokal lanjut.Metode: Desain studi ini adalah kohort retrospektif. Subjek berasal dari pasien KSSRM lokal lanjut yang menjalani kemoterapi induksi di Divisi Bedah Onkologi RSCM periode 2020- 2023). Dilakukan penilaian ADC pada MRI pra pengobatan, selanjutnya dilakukan penilaian respon kemoterapi pada pasien.Hasil: Terdapat 43 subjek dengan nilai median ADC pra pengobatan 1172.49 mm2/s, dan 17 (40%) subjek berespon baik terhadap kemoterapi berbanding 26(60%) subjek tidak respon. Dilakukan analisis hubungan nilai ADC pra pengobatan dengan respon kemoterapi, dengan median nilai ADC pada kelompok responder dibanding non-responder, 1244.04 berbanding 1163.30, nilai p 0.172.Kesimpulan: nilai ADC pra pengobatan tidak berhubungan dengan respon kemoterapi pada kasus KSSRM stadium lokal lanjut.

---

Background: The incidence of oral squamous cell carcinoma (OSCC) has been noted to increase in the last decade, most cases come with locally advanced stages. Induction chemotherapy is a recommendation treatment in most cases of this stage, with purpose that the tumor will shrink so that surgery can be performed better with favourable organ preservation. However, it was noted that 40% of patients did not respond well to induction chemotherapy, most patient would face additional morbidity, tumor progression and worse case became unresectable. The value of the Apparent Diffusion Coefficient (ADC) in Diffusion Weighted - MRI (DW-MRI) is a functional parameter of MRI related to the density of tumor cells and extracellular matrix. So that this examination can be used to predict the patient's chemotherapy response.

Aim: to find association between pretreatment ADC with chemotherapy response of advance stage OSCC patient.

Methods: this is a retrosective cohort study. The subjects are advace stage OSCC patient that undergoing induction chemothrapy at Surgical Oncology Division of Ciptomangukusumo Hospital from 2020 to 2023. Subject’s ADC value was determined at pretreatment MRI and then chemoterapy response was assesed for each subject.

Results: there were 43 subjects, with median pre treatment ADC value was 1172.49 mm2/s, 17 (40%) subjects respond well to chemotherapy and 26(60%) subject were not respond. Further analysis to find association between variables found the median ADC value was 1244.04 for responder subjects vs 1163.30 for non-responder (p 0.172).

Conclusion: pre-treatment ADC value is not associated with chemotherapy response of advance stage OSCC patient."

"Kanker rongga mulut memiliki prevalensi tinggi di banyak negara berkembang, termasuk Indonesia. Diagnosis dini kanker rongga mulut sangat penting untuk meningkatkan peluang keberhasilan pengobatan. Namun, saat ini banyak kasus terdiagnosis pada stadium lanjut akibat keterbatasan akses terhadap metode diagnosis konvensional. Penelitian ini bertujuan merancang model deep learning dengan transfer learning menggunakan arsitektur Convolutional Neural Network, antara lain DenseNet201, EfficientNet-b4, dan InceptionResNetV2 untuk mengklasifikasikan lesi oral menjadi empat kelas: tanpa lesi (healthy), lesi jinak (benign), lesi pra-kanker (OPMD), dan kanker (malignant). Hasil terbaik didapat dari model DenseNet201 dengan penerapan augmentasi data dan focal loss, yang mencapai rata-rata performa akurasi 87.2%, presisi 87.4%, recall 87.4%, dan f1-score 87.2%. Penelitian ini juga mengimplementasikan Explainable AI (XAI) menggunakan metode Grad-CAM dan Score-CAM untuk menghasilkan visualisasi hasil prediksi dan meningkatkan interpretabilitas. Hasil visualisasi menunjukkan bahwa model mampu memfokuskan perhatian pada area lesi di dalam rongga mulut.

---

Oral cancer has a high prevalence in many developing countries, including Indonesia. Early diagnosis of oral cancer is very important to increase the chances of successful treatment, but currently many cases of oral cancer are diagnosed at an advanced stage because of limited access to conventional diagnostic methods. This research aims to develop a deep learning model with transfer learning using Convolutional Neural Network architectures, including DenseNet201, EfficientNet-b4, and InceptionResNetV2 to classify oral lesions into four classes: healthy, benign, Oral Potentially Malignant Disorder (OPMD), and cancer. The best model was achieved by the DenseNet201 model using data augmentation and focal loss, which achieved an average performance of 87.2% accuracy, 87.4% precision, 87.4% recall, and 87.2% f1-score. This research also implemented Explainable AI (XAI) using Grad-CAM and Score-CAM methods to generate visualization of prediction results and improve interpretability. The visualization results show that the model was able to focus its attention on the lesion area in the oral cavity."