Science & Justice
Volume 51, Issue 1 , Pages 24-27, March 2011

Validity of a dichotomous expert response in bitemark analysis using 3-D technology

  • Stella Martin-de-las-Heras

      Affiliations

    • Corresponding Author InformationCorresponding author. Department of Forensic Medicine, School of Medicine, Avenida de Madrid s/n, University of Granada, 18071 Granada, Spain. Tel.: +34 958 249932; fax: +34 958 246107.
    • ,
  • Daniel Tafur

Received 13 April 2010; received in revised form 9 July 2010; accepted 18 July 2010. published online 16 August 2010.

Article Outline

Abstract 

Despite efforts to quantify bitemark evidence, comparison procedures remain subjective and yield different degrees of certainty. Our aim was to study the effectiveness of a comparison procedure requiring a dichotomous response by the expert. We compared overlays from 3-D images of dental casts and bite impressions, obtained using DentalPrint© software. Receiver operating characteristic (ROC) analysis was performed on the results of 104 comparisons, finding an area under the ROC curve of 0.955 (standard error=0.029; 95% CI, 0.896–0.986), sensitivity of 92.3% (95% CI, 74.8–98.8) and specificity of 98.7% (95% CI, 93.0–99.8). According to these findings, this bitemark analysis procedure is highly accurate, although study limitations are discussed, placing these results in context. The main advantage of the dichotomous decision model is that it can be more easily understood, facilitating course of justice. Further research is warranted to explore the potential of this approach as an alternative to diagnostic decisions based on certainty levels.

Keywords: Bitemark, 3-D technology, DentalPrint©, Expert conclusions

 

Back to Article Outline

1. Introduction 

There remains a considerable scope for the further investigation of bitemark analysis. Some researchers have focused on the use of 3-D technology [1], [2], [3], [4], [5], since the dentition of the biter and the corresponding bite marks are both 3-D phenomena. However, comparison of 2-D photographs of the injury with 2-D overlays from the suspect's teeth remains the most common approach [6], [7], [8], [9]. Regardless of whether 2-D or 3-D technology is used to generate comparison overlays from dental casts and bite injuries, the comparison procedure continues to be subjective.

The subjective element of expert evaluations could be minimized by an effective quantitative approach to comparisons. Most authors addressing this challenge have cited the measurable uniqueness of the position of teeth commonly involved in biting [10], [11], [12], [13], [14], [15]. However, the question remains whether these features are detectable in a bitemark left on the skin. Until the difficulties of quantifying dental characteristics of bite injuries on the skin are resolved, qualitative assessments will remain the most appropriate approach to bitemark analysis.

In 1986, the American Board of Forensic Odontology (ABFO) [16] attempted to systematize bite mark analysis by publishing a scoring guide, but the proposal was abandoned due to reports of a wide inter-examiner variability in scores in practice [17], [18]. In 1994, an ABFO workshop produced the Bitemark Terminology Guidelines [19], which establish different degrees of certainty in matching a bitemark to a suspect. As many as 26 terms and their connotations were included in these guidelines, evidencing the lack of precise descriptors in bitemark analysis. ABFO standards are periodically reviewed and subject to revision in the light of developments. The 2009 version of the manual [20] recommends that experts select one of four conclusions in their judgements: “Reasonable Dental/Medical Certainty”, “Probable”, “Exclusion” or “Inconclusive”.

Similar difficulties with expressing the degree of certainty of a match have been reported in other forensic pattern comparisons [21], [22], and it has been speculated that the different conclusion levels are frequently interpreted by courts and jurors as dichotomous, i.e., positive or negative [23].

The present study was prompted by the results of our previous research on the use of 3-D technology (DentalPrint© software) to generate overlays from 3-D images of suspect's dental casts for comparison with 2-D photographic images [4] and 3-D images of indented marks [5]. An ABFO scoring system [16] was used in both investigations, recording the decision as “non-biter”, “possible biter”, “probable biter”, or “biter”. However, analysis of ROC curves in which the conclusions of each examiner served as the cutoff point found no significant differences in sensitivity or specificity among the different cutoff points.

With this background, the objective of the present study was to analyze the effectiveness of a comparison procedure based on a dichotomous expert response. The comparison overlays from dental casts and bite injuries were generated with 3-D technology using freely available software.

Back to Article Outline

2. Material and methods 

2.1. Experimental bitemarks 

Experimental bitemarks were derived from 13 stone models of the dentition of nine adults and four children with varied presence, status, and arrangement of upper and lower anterior teeth. The biting edges of each upper and lower study cast were firmly pressed into an impression tray filled with red wax (Coltène Whaledent®, New York, USA), leaving a shallow imprint (depth of 1–3mm).

Futar® D Fast (Kettenbach GmbH & Co., Eschenburg, Germany) was used to make 5-mm thick reverse models of the wax bite impressions. This polyvinyl siloxane-based addition-curing material sets rapidly (45s) and remains stable and hard (shore-D 43).

2.2. 3-D procedure 

Models and bite impressions were scanned with a 3-D contact-type scanner (Picza 3D Scanner® model PIX-3, Roland DG Corp., Shizuoka, Japan), treating the 3-D images with DentalPrint© software (2004, Department of Forensic Medicine and Forensic Odontology, University of Granada, Spain; freely available at http://www.ugr.es/local/stella/dentalprint).

Comparison overlays of 3-D images of dental casts and bite impressions were produced in three steps, as previously described [2], selecting six upper and six lower anterior teeth for analysis. A contact plane was created from the three highest points on the 3-D images of the dental casts and bite impressions. Finally, biting edges were obtained by using DentalPrint©, which allows the contact plane to extend deep into the teeth. This is a reliable and accurate method for generating comparison overlays [3]; procedures are depicted in Fig. 1.

  • View full-size image.
  • Fig. 1 

    Experimental study of bitemarks: (A) Example of bite impression creation; (B) overlay generation from 3-D image of bite impression using DentalPrint© software; and (C) overlay generation from 3-D image of upper dental cast using DentalPrint© software.

2.3. Comparison process 

3-D overlays from single dental models (upper and lower) were compared with 3-D overlays generated from single sets of bite impressions (upper and lower). The comparison was made by an expert in bitemark analysis, who received a template giving a sequence of one-by-one comparisons between a dentition overlay and an impression overlay, offering a total of 104 comparisons. There were 26 matching dentition and impression overlays. All comparisons were made in a blinded fashion, and the expert did not know how many overlays were likely to yield a positive identification of the biter.

The scoring system used by the expert was dichotomous (biter or non-biter). The “biter” decision was taken when the expert had a reasonable dental medical certainty of a match between dental cast and bitemark, and the “non-biter” decision was taken when this was not the case.

2.4. Statistical analysis 

Receiver operating characteristic (ROC) analysis was used to evaluate the accuracy of the examiner's decisions. The ROC curve combines the concepts of sensitivity and specificity into a single measure of accuracy. Areas under the ROC curve (AUCs) between 0.5 and 1 indicate a positive relationship between the rating scale and correct identification; 95% confidence intervals (CIs) were calculated by using the bootstrap technique.

Back to Article Outline

3. Results and discussion 

This study evaluates a bitemark method in which the expert must make a dichotomous response after comparison of a 3-D image of dentition with a 3-D image of a bitemark. Fig. 2 shows the results of the 104 comparisons. The AUC was 0.955, with standard error of 0.029 and 95% CI of 0.896–0.986, demonstrating, according to Swets' scoring system [24], that this technique is a highly accurate method of bitemark analysis under these conditions. The sensitivity was 92.3% (95% CI, 74.8–98.8), indicating that this approach can be used by an expert examiner to correctly identify the dentition corresponding to a specific bitemark (true positive fraction) and can therefore serve as a screening test. The specificity was 98.7% (95% CI, 93.0–99.8), signifying that an expert examiner can use it to correctly identify dentitions that did not make a specific bitemark (true negative fraction).

  • View full-size image.
  • Fig. 2 

    Receiver operating characteristic (ROC) curves for the dichotomous expert response when comparing the 3-D images of bite impressions and dental casts. AUC, area under the ROC curve; CI, 95% confidence interval; SE, standard error.

The bites were in dental wax, which has been widely used in this research line [3], [5], [25], [26] but cannot accurately reproduce the viscoelastic properties of human skin. It is not possible, using this experimental approach, to consider the influences of skin tension, anatomical location of the bite injury or skin movement during the bite [27]. Our findings should therefore be interpreted with caution, and a lower rate of accuracy could be expected for bitemarks on human antemortem skin, which is a relatively poor impression material in comparison to wax. Another important limitation of our technique is that it is restricted to bite injuries that leave indentations in the skin, known to have a very high forensic significance [28], [29]. By the time an analysis can be performed, bitemark injuries do not usually retain their three-dimensional features.

Although some experts have stated that bitemark analysis has no quantitative basis [30], several attempts have been made to quantify the process [3], [31], [32] and facilitate assessment of the impact of bitemark evidence by judges and juries. However, effective quantification of the comparison procedure and elimination of the subjective element has yet to be achieved. Only one study using 3-D technology has attempted the quantitative comparison of human dentitions with simulated bite marks [3], obtaining true and false positive fractions of only 78% and 15%, respectively, evidencing the need for further research on this complex issue.

Similar accuracy results between assessments based on certainty levels and dichotomous decisions were reported by Pretty and Sweet [23], who used 2-D bitemark overlays and photographs of simulated bites in pig skin. Thus, 10 ABFO Diplomates obtained an AUC of 0.805 when selecting among five levels of certainty (“reasonable medical certainty”, “probable”, “possible”, “exclusion” or “inconclusive”) and an AUC of 0.832 when forced to decide between “biter” and “nonbiter”. A similar level of accuracy (AUC=0.86) was obtained when 32 ABFO Diplomates compared four bitemark cases with seven potential suspects using 2-D overlays [33]; three were authentic cases in which a bitemark was present on the skin of a victim and had been previously investigated and litigated. There were seven possible conclusions (“reasonable medical certainty”, “probable”, “possible”, “improbable”, “incompatible” “inconclusive” or “non-diagnostic”). Hence, despite differences in the substrate and certainty levels, very similar accuracy values were obtained by two studies using ABFO Diplomates and the same bitemark analysis technique.

The present results were very similar to those obtained in the earlier investigation by our group [5] using the same 3-D technology and experimental bitemarks (AUC=0.953), in which the expert selected from four conclusions in accordance with ABFO recommendations [16]. Our results were achieved using impressions in dental wax. In real-life cases, when the injury on the skin may be less clear, this technique may serve to exclude rather than implicate suspected biters. However, the ability to indicate a suspect as a possible biter may be of utility in combination with additional contributing evidence.

The strength of the dichotomous decision model lies in its simplicity and its use of terms (e.g., false positive and true negative) that can be readily and clearly understood by all parties (e.g., judges, juries and other forensic dentists), assisting judicial decision-making. The present findings suggest that the effectiveness of the technique may not be reduced when the expert must decide between biter and non-biter. However, given the current state of the art, experts must continue to report their findings in accordance with ABFO-recommended conclusions. Further research is required into comparative procedures to establish a sound basis (quantitative or qualitative) for diagnostic decisions. In all cases, the expert must explain to the court the technology and experimental bite marks used, the level of expert training, the subjective nature of evaluations and the statistical analysis developed, including error rates.

Back to Article Outline

Acknowledgements 

This work was supported by the Spanish Ministry of Social Affairs National Research Plan (Reference: 101/06).

Back to Article Outline

References 

  1. Thali MJ, Braun M, Markwalder TH, Brueschweiler W, Zollinger U, Malik NJ, et al. Bite mark documentation and analysis: the forensic 3D/CAD supported photogrammetry approach. Forensic Sci. Int. 2003;135:115–121
  2. Martin-de-las-Heras S, Valenzuela A, Ogayar C, Valverde AJ, Torres JC. Computer-based production of comparison overlays from 3D-scanned dental casts for bite mark analysis. J. Forensic Sci. 2005;50:127–133
  3. Blackwell SA, Taylor RV, Gordon I, Ogleby CL, Tanijiri T, Yoshino M, et al. 3-D imaging and quantitative comparison of human dentitions and simulated bite marks. Int. J. Legal Med. 2007;121:9–17
  4. Martin-de-las-Heras S, Valenzuela A, Valverde AJ, Torres JC, Luna-del-Castillo JD. Effectiveness of comparison overlays generated with DentalPrint software in bite mark analysis. J. Forensic Sci. 2007;52:151–156
  5. Martin-de-las-Heras S, Tafur D. Comparison of simulated human dermal bitemarks possessing three-dimensional attributes to suspected biters using a proprietary three-dimensional comparison. Forensic Sci. Int. 2009;190:33–37
  6. Sweet D, Parhar M, Wood RE. Computer-based production of bite mark comparison overlays. J. Forensic Sci. 1998;43:1050–1055
  7. Johansen RJ, Bowers CM. Digital analysis of bite mark evidence. Santa Barbara: Forensic Imaging Services; 1999;
  8. Johansen RJ, Bowers CM. Digital analysis of bite mark evidence using Adobe. Santa Barbara: Forensic Imaging Services; 2000;
  9. Bowers CM, Johansen RJ. Bitemark evidence. In:  Saks MJ editors. Modern Scientific Evidence. New York: West Publishing Co.; 2002;p. 374–426
  10. MacFarlane TW, MacDonald DG, Sutherland DA. Statistical problems in dental identification. J. Forensic Sci. Soc. 1974;14:247–252
  11. Sognnaes RF, Rawson RD, Gratt BM, Nguyen NB. Computer comparison of bitemark patterns in identical twins. J. Am. Dent. Assoc. 1982;105:449–451
  12. Rawson RD, Ommen RK, Kinard G, Johnson J, Yfantis A. Statistical evidence for the individuality of the human dentition. J. Forensic Sci. Soc. 1984;29:245–253
  13. Kieser JA, Bernal V, Wadell JN, Raju S. The uniqueness of the human anterior teeth dentition: a geometric morphometric analysis. J. Forensic Sci. 2007;52:671–677
  14. Kouble RF, Craig GT. A survey of the incidence of missing anterior teeth: potential value in bite mark analysis. Sci. Justice. 2007;47:19–23
  15. Lopez TT, Biazevic MGH, Michel-Crosato E. National survey of the incidence of missing anterior teeth: potential use in bite mark analysis in the Brazilian context. Sci. Justice. December 31 2009;
  16. Guidelines for bite mark analysis, American Board of Forensic Odontology. J. Am. Dent. Assoc. 1986;112:383–386
  17. Vale GL, Rawson RD. Discussion of “reliability of the scoring system of the American Board of Forensic Odontology for human bite mark analysis”. J. Forensic Sci. 1988;33:20
  18. Stimson PG, Mertz CA. Bite mark techniques and terminology. In:  Stimson PG,  Mertz CA editor. Forensic dentistry. Boca Raton: CRC Press; 1997;p. 137–159
  19. ABFO Bite mark methodology guidelines. In:  Bowers CM,  Bell GL editor. Manual of Forensic Odontology. Colorado Spring: American Society of Forensic Odontology; 1995;p. 334–353
  20. Diplomates reference manual. Colorado Spring: American Board of Forensic Odontology; 2009;
  21. Shor Y, Weisner S. A survey on the conclusions drawn on the same footwear marks obtain in actual cases by several experts throughout the world. J. Forensic Sci. 1999;44:380–384
  22. Evett IW, Williams RL. A review of the sixteen points fingerprint standard in England and Wales. J. Forensic Ident. 1996;46:49–73
  23. Pretty IA, Sweet D. Digital bite mark overlays — an analysis of effectiveness. J. Forensic Sci. 2001;46:1385–1391
  24. Swets JA. Measuring the accuracy of diagnostic systems. Science. 1998;240:1285–1293
  25. Nambiar P, Bridges TE, Brown KA. Quantitative forensic evaluation of bite marks with the aid of a shape analysis computer program: Part 2; “SCIP” and bite marks in skin and foodstuffs. J. Forensic Odontostomatol. 1995;13:26–32
  26. Sweet D, Bowers CM. Accuracy of bite mark overlays: a comparison of five common methods to produce exemplars from a suspect's dentition. J. Forensic Sci. 1998;43:362–367
  27. Bush MA, Miller RG, Bush PJ, Dorion RBJ. Biomechanical factors in human dermal bitemarks in a cadaver model. J. Forensic Sci. 2009;54:167–176
  28. Pretty IA. Development and validation of a human bitemark severity and significance scale. In: Proceedings of the AAFS Conference, Seattle. 2006;p. 209
  29. Pretty IA. The barriers to achieving an evidence base for bitemark analysis. Forensic Sci. Int. 2006;159:110–120
  30. Pretty IA, Sweet D. The scientific basis for human bitemark analysis — a critical review. Sci. Justice. 2001;41:85–92
  31. Naru AS, Dykes E. Digital images cross-correlation technique for bite mark investigations. Sci. Justice. 1998;37:251–258
  32. Kittleson JM, Kieser JA, Buckingham D, Herbison GP. Weighing evidence: quantitative measures of the importance of bitemark evidence. J. Forensic Odontostomatol. 2002;20:31–37
  33. Arheart KL, Pretty IA. Results of the 4th ABFO bitemark workshop — 1999. Forensic Sci. Int. 2001;124:104–111

PII: S1355-0306(10)00100-0

doi:10.1016/j.scijus.2010.07.003

Science & Justice
Volume 51, Issue 1 , Pages 24-27, March 2011

Article Tools

* Image Usage & Resolution