Prevalence and imaging characteristics of detectable tonsilloliths on 482 pairs of consecutive CT and panoramic radiographs

Of the 482 individuals, 222 (46.1%) were judged to have tonsilloliths on CT scans (Table 1). The 222 patients with tonsilloliths on CT images consisted of 91 males and 131 females, and no difference by sex in the tonsillolith detection rate was seen. However, a significant correlation was found between age and the tonsillolith detection rate (Spearman r = 0.176, p < 0.0001). As the subjects aged, the detection rate increased gradually. In particular, a significant difference in the tonsillolith detection rate was found between the over and under 40-year-old groups (p < 0.0001), but the tonsillolith detection rate was relatively stable among the various groups (Table 2). Overall, 99 patients had one calculus, 37 had two, 27 had three, 21 had four, 12 had five, and so on (Table 3), up to a maximum of 28 (Figure 1). The shapes of tonsilloliths varied, including dot, round, oval, and irregular structures (Figure 2). Round or oval tonsilloliths with a long diameter under 2 mm were considered dots, while those with a long: short diameter ratio of under 3:2 were considered round, those with a long: short diameter ratio of over 3:2 were considered oval, and those with shapes that were neither round nor oval were considered irregular. The most common shape was the dot lump and/or lumps of structure(s) with high CT number(s) within the palatine tonsil on CT images. With respect to the sizes of the tonsilloliths, 59 were under 2 mm, 149 were over 2 mm and under 5 mm, 14 were over 5 mm and under 10 mm, and none were over 8 mm (Table 4). The most common length was from under 3 to 4 mm on CT images. On CT, 63 (13.1%) were on the right side, 78 (16.3%) were on the left side, and 81 (16.8%) were bilateral (Table 5); the rates were almost equal, with no significant difference observed (p = 0.194). Furthermore, 157 were under the soft palate, 95 were coincident with the soft palate, and 0 were over the soft palate. At the same time, 95 were in the medial half and the others were in the lateral half of the tonsil. The calcification levels of tonsilloliths on CT images were over 100 HU and under 300 HU in 95, over 300 HU and under 500 HU in 68, over 500 HU and under 1000 HU in 46, and over 1000 HU in 13 (Table 6).

Of the 482 individuals, 37 (7.7%) (Figure 3) were judged to have tonsilloliths on panoramic radiographs, but 222 (46.1%) were judged to have tonsilloliths on CT scans (Table 1). A significant difference in the tonsillolith detection rate was found between the two modalities (p < 0.0001). Of the 37 patients judged to have tonsilloliths on panoramic radiographs, 2 did not actually have calcifications on CT images (Figure 4). In both cases, enostosis in the mandibular ramus was detected on CT images (Figure 4). Therefore, there were actually 35 (7.3%) patients with tonsilloliths on panoramic radiographs. Almost all tonsilloliths showed a radiopaque nodular mass or masses on panoramic radiographs. The size on panoramic radiographs was over 2 mm and under 5 mm in 29, over 5 mm and under 10 mm in 6, and over 10 mm in 0 (Table 4). Fourteen were located on the right side, 18 were on the left side, and 3 were bilateral (Table 5). At the same time, 25 were piled up on the mandibular ramus and under the soft palate, 0 were on the mandibular ramus and over the palate, and 10 were on the mandibular ramus and coincident with the soft palate on panoramic radiographs.

There were 185 cases of disagreement between panoramic radiographic and CT imaging findings (Table 1). It was presumed that CT numbers and tonsillolith size might explain why the tonsilloliths were detected on CT images, but not on panoramic radiographs. Certainly, there was a significant correlation between CT number and the tonsillolith detection rate on panoramic radiographs (Spearman r = 0.429, p < 0.0001). The higher the CT number was, the higher the tonsillolith detection rate was. All 95 cases with CT number < 300 HU were not detected on panoramic radiographs (Table 6). Furthermore, a significant correlation was observed between size and tonsillolith detection rate on panoramic radiographs (Spearman r = 0.318, p < 0.0001). The tonsillolith detection rate was higher as size increased. In fact, in 179 of 187 cases that tonsilloliths were not detected on panoramic radiographs, calculus sizes were <5 mm (Table 4). In addition, a significant correlation was found between the tonsillolith detection rate on panoramic radiographs and the number of tonsilloliths (Spearman r = 0.333, p < 0.0001). However, in three cases, the two main causes mentioned above did not appear responsible (Figure 5). In all cases, the tonsilloliths were at a relatively low level that coincided with the mandibular ramus and the head of the submandibular glands. These cases shared no other similar characteristics (Table 7).