Evaluation and comparison of saw mark features using a stereomicroscope and micro-XCT scanner

Abstract

Saw mark analysis is a vital component of forensic anthropology, particularly in the investigation of post-mortem dismemberments. Despite its medicolegal significance, this field remains underdeveloped in South Africa, where a high incidence of murder, including dismemberments, occurs annually. The scarcity of trained practitioners capable of analysing saw marks often results in the loss of critical evidence regarding the tools used in these crimes. This study aimed to evaluate and compare saw mark features created by a 6 tooth per inch (TPI) crosscut saw on human femora using a stereomicroscope and micro-focus X-ray computed tomography (micro-XCT) modality. It also sought to validate the use of three-dimensional (3D) printed saw mark replicas for courtroom presentations and educational purposes. By addressing these objectives, the research advances saw mark analysis methodologies and their practical applications in forensic contexts. Four fresh human femora were obtained from the National Tissue Bank of South Africa, defleshed, and sectioned. Thirty-six bone sections were prepared, yielding 32 false starts and 64 complete cuts. Analyses were conducted using a stereomicroscope, micro-XCT, and 3D print modality, assessing 16 unique saw mark features. Statistical evaluations included reproducibility assessments, modality comparisons, and validation of 3D printed models. Intra- and inter-observer error rates were calculated using Cohen’s Kappa, proportion of agreement, and technical error of measurement. Comparative analyses were performed using Chi-Squared tests, Paired Student’s t-tests, and Wilcoxon signed-rank tests. Reproducibility assessments demonstrated high intra- and inter-observer agreement for most false start and complete cut features, although agreement varied by modality. Some features, such as minimum kerf width (MKW), kerf wall shape, and kerf floor shape, showed perfect or near-perfect agreement across all modalities. In contrast, features such as pull-out striae, kerf flare, and entrance shaving exhibited lower agreement, particularly in 3D print evaluations. Measurement reproducibility was highest for the micro-XCT modality, followed by the stereomicroscope and 3D prints modalities, with intra-observer measurements yielding lower error rates than inter-observer measurements. Modality analysis revealed that stereomicroscope and micro-XCT analysis offered complementary strengths, providing detailed visualisation of nuanced features, whereas 3D prints were less reliable for capturing fine details. Statistical validation highlighted the limited comparability between 3D prints and stereomicroscopic analysis of complex features. Stereomicroscope analysis emerged as the most effective modality for detecting detailed saw mark features, followed by micro-XCT, whereas 3D prints demonstrated limitations in replicating subtle textures. Minimum kerf width was consistently measurable across all modalities, but complex features, such as exit chipping and exit shaving, showed better agreement between the stereomicroscope and micro-XCT modalities than with 3D prints. Significant differences across modalities underscore the importance of using analytical methods in obtaining reliable results. While 3D prints have the potential for teaching and demonstrations, they require further refinement to accurately replicate intricate saw mark features. Future research should focus on standardising the measurement methodologies for complex features to improve reproducibility and consistency. Higher resolution micro-XCT scanning may enhance the non-destructive visualisation of fine details. Additionally, studies on a broader range of saw types could deepen the understanding of feature variations across tool classes and modalities, thereby strengthening forensic applications.