ABSTRACT
BACKGROUND: Locally advanced rectal cancer often requires neoadjuvant treatment (NAT) before surgical intervention. This study aimed to assess the concordance between preoperative magnetic resonance imaging (MRI) findings and postoperative pathology results after NAT in rectal cancer patients. METHOD: A retrospective analysis of 52 patients who underwent NAT and subsequent surgery at Ankara Bilkent City Hospital between May 2019 and May 2023 was conducted. Demographics, preoperative MRIs, time intervals between NAT, MRI, and surgery, and postoperative pathology were assessed. RESULTS: The median age of the cohort was 59 years, with a male predominance (76.9%). Tumour T stage (κ = 0.157), lymph node stage (κ = 0.138), and circumferential resection margin (κ = 0.138) concordance showed poor agreement between post-neoadjuvant treatment (PNT) MRI and pathology. PNT MRI demonstrated a limited correlation with postoperative pathology. CONCLUSIONS: While preoperative MRI is commonly used for restaging after NAT in rectal cancer, our study highlights its limited concordance with postoperative pathology. The sensitivity and specificity metrics, although reported in the literature, should be interpreted alongside concordance assessments for a comprehensive evaluation.
Subject(s)
Magnetic Resonance Imaging , Neoadjuvant Therapy , Rectal Neoplasms , Humans , Rectal Neoplasms/pathology , Rectal Neoplasms/diagnostic imaging , Rectal Neoplasms/surgery , Rectal Neoplasms/therapy , Neoadjuvant Therapy/methods , Male , Magnetic Resonance Imaging/methods , Middle Aged , Female , Retrospective Studies , Aged , Neoplasm Staging , Preoperative Care/methods , AdultABSTRACT
This study focuses on the geo-statistical assessment of spatial estimation models in forest crimes. Used widely in the assessment of crime and crime-dependent variables, geographic information system (GIS) helps the detection of forest crimes in rural regions. In this study, forest crimes (forest encroachment, illegal use, illegal timber logging, etc.) are assessed holistically and modeling was performed with ten different independent variables in GIS environment. The research areas are three Forest Enterprise Chiefs (Baskonus, Cinarpinar, and Hartlap) affiliated to Kahramanmaras Forest Regional Directorate in Kahramanmaras. An estimation model was designed using ordinary least squares (OLS) and geographically weighted regression (GWR) methods, which are often used in spatial association. Three different models were proposed in order to increase the accuracy of the estimation model. The use of variables with a variance inflation factor (VIF) value of lower than 7.5 in Model I and lower than 4 in Model II and dependent variables with significant robust probability values in Model III are associated with forest crimes. Afterwards, the model with the lowest corrected Akaike Information Criterion (AICc), and the highest R2 value was selected as the comparison criterion. Consequently, Model III proved to be more accurate compared to other models. For Model III, while AICc was 328,491 and R2was 0.634 for OLS-3 model, AICc was 318,489 and R2 was 0.741 for GWR-3 model. In this respect, the uses of GIS for combating forest crimes provide different scenarios and tangible information that will help take political and strategic measures.
