Discussion

Firearm-related violence continues to represent a major global public health and security concern and accounts for a substantial proportion of homicides worldwide [16, 37]. Within this context, forensic ballistics plays a critical role in reconstructing shooting events and establishing associations among suspects, victims, and firearms [8, 15, 38]. The analysis of gunshot residue (GSR) constitutes one of the most informative trace-based approaches within this field, as the presence, distribution, and composition of residue particles may support inferences regarding discharge occurrence, firing distance, and potential secondary transfer [18, 39–44].

GSR is composed of heterogeneous inorganic and organic fractions generated during cartridge ignition, including lead, barium, and antimony-rich particles that are relatively stable and radiodense, as well as more volatile organic residues [45, 46]. Because these particles are microscopic and irregularly distributed, analytical techniques capable of non-destructive detection, three-dimensional localization, and morphometric characterization are particularly advantageous [18, 42, 46, 47]. Within this framework, micro-CT has emerged as a complementary modality that enables volumetric visualization and quantification of radiopaque microparticles across diverse substrates [22]. Owing to its reliance on differences in X-ray attenuation coefficients expressed in Hounsfield units, micro-CT preferentially detects the inorganic, metal-rich fraction of GSR, which produces sufficient contrast for reliable identification [47, 48].

Despite these theoretical advantages, the literature has lacked a quantitative synthesis evaluating the performance of micro-CT for GSR detection and its potential application to shooting distance estimation [11, 31–36]. The present systematic review therefore sought to consolidate and critically appraise the available experimental evidence in both biological and non-biological matrices. The pooled analyses demonstrated a consistent decline in micro-CT–detectable GSR with increasing shooting distance. This inverse association was observed across all modeling strategies and aligns with the expected reduction in particle deposition density as dispersion increases.

While between-study heterogeneity was substantial, this variability likely reflects the intrinsically multifactorial nature of GSR deposition, including differences in substrate characteristics, particle behavior, local deposition dynamics, and experimental conditions. Importantly, inorganic residues remained detectable even under heterogeneous conditions, which supports the feasibility of micro-CT as a screening and localization tool. However, the magnitude of this heterogeneity also limits the interpretability of pooled estimates, which cannot be interpreted as representing a single underlying effect.

Distance-stratified meta-analyses further illustrated this pattern. Deposition levels were highest at short range (5 cm), followed by a marked reduction at intermediate distances (15 cm), and low but still measurable detection at 30 cm. These findings are consistent with classical ballistic observations indicating that particle concentration decreases rapidly with distance but that heavier fragments may still reach more distal targets. Persistent heterogeneity across strata reinforces that residue retention depends not only on distance but also on substrate texture, porosity, and surface chemistry.

The quadratic mixed-effects meta-regression provided additional evidence by modeling the continuous relationship between distance and detection probability. The significant quadratic component confirmed a nonlinear decay pattern, with an estimated vertex at approximately 17.8 cm. This value should not be interpreted as a universal physical threshold but rather as a dataset-specific summary reflecting the combined influence of emission dynamics, particle inertia, and experimental conditions. The pseudo-R² of 0% indicates that the variability observed across studies was not meaningfully explained by shooting distance alone, reinforcing the multifactorial nature of GSR deposition and the influence of additional factors such as substrate properties and local deposition dynamics.

The observed nonlinear decay in detection probability is consistent with the underlying physical mechanisms governing GSR generation and transport [49]. During discharge, primer detonation and propellant combustion produce a heterogeneous spectrum of particles varying in mass, density, and thermal state, including metallic fragments, condensed vapors, and agglomerates [18]. These particles are subsequently transported by the combined effects of the expanding gas plume, air resistance, gravity, and local turbulence [47, 50]. Consequently, dispersion is not strictly monotonic but results from the interaction between ballistic inertia and aerodynamic forces.

Heavier and denser particles exhibit greater momentum and tend to maintain trajectories closer to the projectile axis, preferentially depositing along the line of fire [34, 36, 51]. Under favorable conditions, such particles may reach intermediate or even relatively distant targets with limited angular deviation. In contrast, lighter particles are readily deflected by muzzle turbulence, vortical flows, and environmental air currents, producing broader and less predictable spatial distributions [18]. Additional factors, including firearm design, barrel length, ammunition composition, gas dynamics, and environmental conditions, further modulate these patterns, explaining why deposition profiles vary substantially across experimental setups [18, 52–54]. This physical complexity provides a mechanistic explanation for both the curved distance–detection relationship and the substantial between-study heterogeneity observed in the meta-analyses.

Within this framework, Gaussian functions offer a reasonable phenomenological approximation of deposition behavior, capturing a central high-concentration region surrounded by progressive stochastic dispersion [31, 51]. The apex of these curves corresponds to the point of maximum particle deposition and has been reported to range from a few centimeters to distances exceeding 10 m, depending on firearm characteristics, ammunition type, gas dynamics, target properties, and environmental conditions [51, 55, 56]. The fitted curves obtained in the present study yielded substrate-specific profiles that reflected differences in particle retention and detectability. Fresh biological tissues exhibited higher initial loads followed by rapid decay with increasing distance, whereas decomposed or charred substrates showed attenuated and more uncertain profiles. Synthetic materials frequently demonstrated reduced retention and flatter curves, suggesting limited adherence of microparticles. These findings indicate that detectability depends not only on emission and transport dynamics but also on the physicochemical properties of the receiving surface, including porosity, roughness, moisture content, and structural integrity.

The interpretative value of these Gaussian models varied substantially according to substrate type. High R² values for fresh biological tissue and cotton indicate that Gaussian functions adequately described the data under relatively homogeneous surface conditions, while negative R² values observed for substrates such as jeans, leather, and decomposed tissues indicate no improvement relative to a constant mean baseline or a null (mean-only) model. These findings suggest that the observed signal is largely dominated by stochastic variability and substrate-related effects, including surface porosity, texture-dependent particle loss, and structural degradation. In these cases, the Gaussian formulation represents only a limited descriptive approximation and underscores the difficulty of modeling residue behavior across heterogeneous forensic matrices.

From a practical forensic perspective, these results support the role of micro-CT primarily as a non-destructive screening and localization technique. The method enables rapid three-dimensional mapping and segmentation of radiodense microparticles, which may guide targeted confirmatory analyses and preserve sample integrity for subsequent testing. However, because image contrast is based exclusively on X-ray attenuation, micro-CT does not provide elemental specificity. Radiodense voxels may correspond to GSR particles but may also represent unrelated dense contaminants or environmental debris. Moreover, although documented contamination or secondary transfer events are relatively uncommon, the possibility of exogenous particle deposition further underscores that detected residues must always be interpreted within the broader ballistic and situational context rather than in isolation [43, 44].

In operational terms, micro-CT should be understood as an additional tool within the forensic workflow and not as a confirmatory or stand-alone method for GSR identification. Its primary utility lies in the rapid volumetric detection of radiodense structures (typically above 1000 HU), enabling the identification of regions of interest that may contain potential residue accumulations. Thus, micro-CT functions as a triage tool for analytical prioritization, supporting the selection of targeted areas for subsequent SEM/EDS examination. However, this sensitivity is not accompanied by chemical specificity, and therefore the method is inherently prone to non-specific detection of other high-density materials, including environmental contaminants and structurally similar artefacts. As a consequence, micro-CT may generate false-positive signals when interpreted in isolation. Accordingly, its role is restricted to preliminary localization and prioritization of suspect regions for subsequent analytical validation, without allowing direct inference of GSR presence. For these reasons, micro-CT findings should be interpreted alongside confirmatory analytical methods, particularly SEM/EDS, when definitive elemental characterization is required.

The principal limitation of this review is the limited and methodologically homogeneous evidence base. Only six studies satisfied the eligibility criteria, most with small sample sizes and substantial imbalance in design. Notably, five originated from the same research group and relied on highly similar experimental conditions, including firearms, calibers, ammunition lots, and micro-CT acquisition protocols. This concentration of evidence within a single research group affects the interpretation of the findings in several ways. It limits external validity, reducing extrapolation to other firearms, ammunition types, substrates, and environmental conditions. It also reduces the independence of observations across studies, since shared protocols and experimental assumptions may produce correlated outcomes. In addition, study-specific characteristics may have exerted a stronger influence on the observed effects than broader generalizable patterns. Therefore, the available evidence is better understood as reflecting a single-laboratory framework instead of an independent body of literature. Consequently, the pooled estimates derived from the meta-analyses, meta-regression, and Gaussian modeling should be interpreted primarily as descriptors of this constrained setting, not as broadly applicable parameters. The absence of independent studies from other research groups further reinforces this limitation.

An additional limitation concerns the uneven distribution of shooting distances across studies. The 23 and 40 cm conditions were represented by only a single study each, limiting the statistical representation of these distances within the pooled dataset. Sensitivity analyses demonstrated that exclusion of these conditions did not substantially alter the pooled estimates or statistical significance of the overall meta-analysis, supporting the robustness of the aggregated findings. Nevertheless, the limited availability of independent observations at intermediate and longer distances restricts the precision of continuous modeling approaches and highlights the importance of additional studies with broader and more balanced distance distributions.

Despite these limitations, the quantitative syntheses and modeling approaches consistently demonstrate that micro-CT is capable of detecting inorganic GSR particles across a range of substrates and distances, with detection probability decreasing as shooting distance increases. The quadratic meta-regression provides an overall representation of this nonlinear relationship within the analyzed dataset, whereas stratified and Gaussian models highlight the influence of substrate-specific retention mechanisms.

The extremely high residual heterogeneity (I² > 99%) indicates that the aggregated estimates were strongly influenced by methodological and structural variability and should not be interpreted as representing a single consistent effect across studies. Consequently, summary measures must be interpreted with caution, as they primarily reflect the dispersion of study-specific scenarios instead of a stable overall relationship. These findings further indicate that shooting distance alone explains little of the observed variability, emphasizing the multifactorial behavior of GSR deposition.