Purpose: Mandibular third molar extraction is a commonly performed procedure and is recognized as a relatively frequent cause of inferior alveolar nerve (IAN) injury. The aim of the present study was to investigate the specific risk factors for neurosensory deficits, including age, gender, impaction depth, angulation of the third molar, and various radiographic superimposition signs. Materials and Methods: In a case-control study of patients who had undergone mandibular third molar extraction, a case group was developed of patients showing neurosensory deficits of the IAN, and a control group was formed of randomly selected patients without any neurosensory symptoms. Bivariate analyses were performed to assess the relationship between each variable and IAN injury. A multivariate logistic regression model was used to compute the odds ratios, P values, and predictive values of the radiographic superimposition signs. Results: Of 12,842 total patients, the study group included 104 cases and 135 controls. The results indicated that older age and deeper impaction status were significant risk factors (P <.05). Darkening of the roots, deflection of the roots, narrowing of the roots, dark and bifid apexes of the roots, and narrowing of the canal were also significant risk factors. The positive predictive values ranged from 0.7% to 6.9% and the negative predictive values from 99% to 100%, with adjustment for the definitive prevalence of IAN injury (0.81%, 104/12,842 patients). However, the relatively low positive predictive value renders questionable the predictability of superimposition signs on orthopantomography. In the absence of specific radiographic signs, the risk of neurosensory deficit of the IAN could be negligible. The sensory symptoms disappeared after 6 months in 92.3% of the patients and 98.1% showed recovery after 1 year. Conclusions: The results of the present study have demonstrated a significant association between several risk factors and neurosensory deficits of the IAN after third molar extraction. With a case group of 104 patients, the number of subjects was significantly greater than that in previous studies, increasing the reliability of these results.