Standardized measurements of radiographic films of the frontal sinuses: An aid to identifying unknown persons.

Abstract

Many methods have been used to identify unknown persons. The most common is the study of fingerprints, which is one of the few methods in which data can be stored and retrieved in a precise and cost-effective manner.

The asymmetry of the frontal sinuses has stimulated several attempts to identify persons by analyzing measurements of the sinuses obtained from plain x-ray films. However, an objective and reproducible method has not been achieved.

The objective of this paper is to report on a method of making standardized measurements of the frontal sinuses on radiographs and storing the resultant information in a computer databank. To test the system, the author made new measurements on 100 x-rays and entered the results into a computer databank that contained information on earlier measurements of the same 100 x-rays and on 400 others. With the aid of the databank, the author was able to rapidly and accurately match each new measurement with the earlier measurement. The author concludes that such a system is a simple and cost-effective method of establishing the identification of unknown persons for whom frontal sinus x-rays exist.

Introduction

Zuckerkandl in 1895 was the first scientist to call attention to the asymmetry of the frontal sinus. [1] His finding was confirmed by many others who performed anatomy studies by dissection and/or radiography: Sieur and Jacob in 1901, [2] Hajek in 1926, [3] Testut and Jacob in 1929, [4] Segura et al in 1943, [5] and Arauz in 1 943. [6] All these authors agreed that the frontal sinus stops growing near the age of 20 years and remains stable throughout the remainder of life. Some factors can modify the normal anatomy of the frontal sinus--fractures, neoplasias, severe infections, and mucoceles--but such occurrences are rare.

Schuller was the first author to note that no two persons have identical frontal sinuses, including identical twins. [7,8] Culbert and Law, [9] Guthrie and Scott, [10] Marek et al, [11] Yoshino et al, [12] and others reported the same findings. It was Schuller who first suggested the possibility of identifying persons by comparing sinus radiographs. [8]

In 1927, Culbert and Law reported that they made the first such identification of a person to be accepted in an American court . [9] Similar reports by Yoshino et al, [12] Atkins and Potsaid, [13] and Jablonski and Shum [14] followed. Kullman et al reported on a study of frontal sinus radiographs of 99 persons that were taken on two separate occasions at various intervals of time. [15] In separate analyses, three observers were able to identify each person by matching the two radiographs with 100% accuracy.

From these reports, there seems to be no doubt that frontal sinus radiographs are a valid aid to identification. Of course, such identification can been made only if the investigator is able to find a previous x-ray of a known person that can be compared with a new x-ray of an unknown person. Locating a matching film, if one even exists, from among a vast number of possibilities can be a cumbersome, time-consuming, and daunting task. A method is needed that would make it possible to file sinus-measurement data from a great number of known persons that could be easily retrieved for later comparison with new films of unidentified persons. According to Lopes, Delclos made the first such attempt in 1934 while working on his doctoral dissertation. [16] Other studies followed, including those by Schuller, [8] Marek et al, [11] and Yoshino et al. [12]

However, even with a good system for storing and retrieving sinus-measurement data, identification is still hampered by the lack of a standardized method of measuring the many variables one encounters--among them anatomic asymmetry, variations in the angle at which x-rays are received, and the presence of incomplete lobulations, nondefined intersinus septa, and accessory septa. Because the data that do exist were all gathered by different researchers, the inevitably subjective nature of the measurements precludes making valid comparisons.

The purpose of this paper is to describe a practical method of making standard measurements of frontal sinus radiographs in humans that can be filed for later comparison with new radiographs of unidentified persons.

Four primary measurements

Measuring normal anatomy. The first step in the measurement of the frontal sinus is to place the x-ray film on a viewer and draw a line directly on the film horizontally along the upper limit of both orbital cavities (figure 1). This is the baseline. Four more lines are drawn perpendicular to the baseline at specific points. One line (figure 1, solid line E) delineates the maximum lateral limit of the right frontal sinus. Another line (F) passes through the highest point (the most distant from the baseline) of the right frontal sinus. A third line (G) is drawn through the highest point of the left frontal sinus. The fourth line (H) defines the lateral limit of the left frontal sinus. From these lines, the four primary measurements (measurements A-D) can be obtained (figure 1, dotted lines):

Measurement A is the diameter of the frontal sinuses at the widest point--that is, the distance between the projected lines that delineate the maximum lateral limits of the right and left sinuses (line E to H).

Measurement B is the distance between the projected lines marking the highest points of the right and left sinuses (line F to G).

Measurement C is the distance between the lines marking the maximum lateral limit and the highest point of the right frontal sinus (line E to F).

Measurement D is the distance between the lines marking the maximum lateral limit and the highest point of the left frontal sinus (line G to H).

Because this technique measures the distance between projected, parallel lines, it corrects any possible distortion that might have been caused by improper positioning (longitudinal) of the skull at the moment the x-ray was taken.

Measuring variations in anatomy. Measuring anatomy that does not fit a classic architectural pattern requires adapting the methods explained above. To facilitate this, the author has developed nine rules to follow:

1) Take measurements only of the air-containing cavities of the frontal sinuses.

2) When one sinus has two equally high points, measure the one closer to the intersinus septum (figure 2, line F).

3) When the highest point is difficult to determine because the sinus has an open-curve lobulation, measure the point at the middle of the lobulation (figure 3, line F).

4) When the highest point is not evident because of a plateau lobulation, measure the middle of the plateau (figure 4, line F).

5) Any air-containing cavity of the frontal bone is considered to be part of the frontal sinus and should be measured (figure 5, lines E, H).

6) When the frontal sinus is triangular in shape and there are no two distinct highest points because they coincide at the vertex of the triangle, measure and notate the highest point of the vertex and list measurement B as zero (figure 6).

7) When the points delineated by lines E, F, G, and H are above the baseline, they are classified as positive (p); when they are below the baseline, they are classified as negative (n). Points are classified as null (0) when they are on the baseline (figure 7).

8) If a skull has only one frontal sinus, make measurement A the distance between the projections of its lateral and medial limits (figure 8). Denote measurement B as zero. Measurement C remains the same. Make measurement D the distance between the highest point and the medial limit.

9) If a skull has no frontal sinuses, measure only the distance between the projection of the lines that pass through the medial borders of the orbit cavities (figure 9, measurement X).

Testing the system

To test the accuracy of this measurement system, 500 plain radiographic films of the frontal sinuses taken in the frontal nasal position were chosen at random from a pool of radiographic films in the outpatient files in the department of otolaryngology at a large community hospital. All radiographs were of adults who had no signs of frontal sinus disease. These films were numbered from 1 to 500.

The frontal sinuses on the 500 x-rays were measured in accordance with the method described. Measurements of all distances A, B, C, and D (and occasionally X) and the characteristics of all lines E, F, G, and H were notated. Data on all A, B, and C measurements were entered into a personal computer according to their lengths, from the shortest to the longest beginning with measurement A. When more than one measurement A had the same length, the data were then ordered according to the length of measurement B. When measurements A and B were identical, the computer ordered them according to the length of measurement C. (The computer could not accept D measurements in this fashion.)

Next, 100 x-rays were randomly chosen from the original set of 500. The 100 films contained a variety of all types of frontal sinuses. The original measurement lines that had been drawn on the radiographs were erased with cotton balls moist with alcohol. All identification was removed from the 100 films. New measurements were made, and the new data were entered into the computer.

The computer was then instructed to search for matches. The new measurements (in cm) were analyzed by the computer and compared with those in the database. To account for the inevitable subjectivity of manual measuring and the consequent variations that can occur from one measurement to another on the same film, each measurement was entered as a range of numbers ([plus or minus]2 mm) rather than as a single figure. For example, a given measurement x would be entered into the computer as being longer than x - 3 and shorter than x + 3. Measurements A, B, C, and D were studied in this manner. The characteristics of lines E, F, G, and H--including the p, n, and 0 classifications indicating their location relative to the baseline--were considered only later to complete the identification.

When the computer found more than one film with identical measurements A, B, C, and D or lines E, F, G, and H, these matches were compared visually. We were able to retrieve films from the database quickly and compare them visually with 100% accuracy.

Discussion

The method described here can be useful in identifying persons who are at risk of being killed or maimed by explosive violence that results in gross disfigurement-- persons such as soldiers, members of flight crews, policemen, and firemen. Having their frontal sinus x-ray measurements on file would assist in later identification if other methods were to fail or be unavailable. Frontal sinus measurements could also quickly identify long-decayed corpses, and they are less expensive than DNA tests.

This preliminary study demonstrates that the system described here is feasible. The author is currently working to make it more effective. The system architecture is based on digitized hardware. A prototype version has been installed in VisualBasic for Windows and is currently available. Software has been fully tested.

The primary advantage of this system is that it allows the computer to narrow the field of possible matches to just a few at most. Even so, one must always bear in mind that the system is only an aid, and that any final identification will be made by an expert's visual comparison of x-ray films.

From the Department of Otolaryngology, Santa Casa School of Medicine of Sao Paulo, Brazil.

Reprint requests: Fernando de Andrade Quintanilha Ribeiro, Santa Casa de Misericordia de Sao Paulo, Departmento de Otorrinolaringologia, Rua Dr. Cesario Motta Junior, 112, Vila Buarque Cep: 01221-020, Sao Paulo-SP Brazil. Phone/fax: [phone]; e-mail: [e-mail]

References

(1.) Zuckerkandl E. Anatomie des sinus frontaux. In: Anatomie Normale et Pathologique des Fosses Nasales et de Leurs Annexes Pneumatiques. Paris: G. Masson, 1895:349-61.

(2.) Sieur and Jacob. Des sinus frontaux. In: Recherches Anatomiques, Cliniques et Operatoires sur les Fosses Nasales et Leur Sinus. Paris: J. Rueff, 1901:83-91.

(3.) Hajek M. Normal anatomy of the frontal sinus. In: Pathology and Treatment of the Inflammatory Diseases of the Nasal Accessory Sinuses. St. Louis: C.V. Mosby, 1926:35-43.

(4.) Testut L, Jacob O. Sinus frontaux. In: Traite d'Anatomie Topographique avec Applications Medico-Chirurgicales. Paris: Librarie Octave Doin, Gaston Doin & Cie., 1929:29-33.

(5.) Segura EV, Canuyt G, Errecart PL, Del Carril AV. Enfermidades de los senos paranasales. In: Otorrinolaringologia Pratica. Buenos Aires: Libreria Hachette, 1943:601-2.

(6.) Arauz SL. Seno Frontal. Buenos Aires: El Ateneo, 1943:11-7.

(7.) Schuller A. Das roentgenogramm der Stirnhohle: Ein hilfsmittel fur die identitatsbestimmung von schadeln. Monatschrift fur Ohrenh 1921;55:1617-20

(8.) Schuller A. A note on the identification of skulls by x-ray pictures of the frontal sinuses. Med J Aust 1943;1:554-7.

(9.) Culbert WL, Law FM. Identification by comparison of roentgenograms of nasal accessory sinuses and mastoid processes. JAMA 1927;88:1634-6.

(10.) Guthrie D, Scott CE. Anatomy: External nose, the nasal cavity, the nasopharynx and the paranasal sinuses. In: Diseases of the Nose, Throat and Ear. Baltimore: William Wood, 1936:1-19.

(11.) MarekZ, Kusmiderski J, Lisowski Z. Radiogramme der stirnhohlen als grundlage fur die identifizierung von katastrophenofern und an unbekannten skeletten. Arch Kriminol 1982;172:1-6.

(12.) Yoshino M, Miyasaka S, Sato H, Sets S. Classification system of frontal sinus patterns by radiography: Its application to identification of unknown skeletal remains. Forensic Sci Int. 1987;34:289-99.

(13.) Atkins L, Potsaid MS. Roentgenographic identification of human remains. JAMA 1978;240:2307-8.

(14.) Jablooski NG, Shum BS. Identification of unknown human remains by comparison of antemortem and postmortem radiographs. Forensic Sci Int 1989;42:221-30.

(15.) Kullman L, Eklund B, Grundin R. Value of the frontal sinus in the identification of unknown persons. J Forensic Odontostomatol 1990;8:3-10.

(16.) Lopes L. Identificacion radioscopica y radiografica. In: Teenica Medico-Legal, Criminalistica. Valencia: Saber, 1953:278-82.