Concordance between clinical and histopathologic diagnosis and an audit of oral histopathology service at a Nigerian tertiary hospital

Introduction Although histopathological diagnosis remains the gold standard; good clinical impression is potentially a key diagnostic tool in rural resource-limited settings. Thus, good concordance between clinical impression and histopathological diagnosis is thus a very crucial diagnostic oral pathology tool in low- and middle-income countries (LMICs). Methods This retrospective study was performed at the Oral pathology and Oral diagnoses units of Obafemi Awolowo University Teaching Hospital Complex (OAUTHC), Osun state. Clinicopathological reports of all biopsies between the period of 2008-2017 were retrieved and analyzed. Descriptive analysis of data was done using Stata 14. Frequency of oral lesions and rates of accurate clinical diagnoses were evaluated for lesional sites and clinician's qualification/specialization. Results In 592 biopsied cases, the mean age was 36.1years with higher female predilection (54.4%). Odontogenic tumors (OTs) were the most prevalent category of lesions (25.3%, n=149), followed by reactive lesions (12%, n=71). Absolute concordance was recorded for 54.6% (k=0.5) of the cases; with highest concordance observed in fibro-osseous lesions (65.6%, k=0.43), and least in pulp/periapical lesions (3.5%). Concordance was higher in females (59.5%, k=0.53) than males (48.3%, k=0.44). Oral medicine specialists had the highest concordance index (62.5%, k=0.59). Conclusion The findings in this research indicate that, on a general note, the degree of concordance between clinical and histopathological diagnosis is poor. Hence, improvement in diagnostic skills (irrespective of clinical specialty) is important to improve treatment outcomes, particularly in LMICs. Continuous personnel training and utilization of advanced diagnostic techniques can potentially help bridge the diagnostic gaps.

. Therefore, a wrong clinical impression might lead to inappropriate investigation; hence, leading to diagnostic pitfalls, delay in management or inappropriate patient management protocols.
Varying degree and percentage of discordance between clinical impressions and histopathological diagnoses has been reported in scientific literature, many of which are attributed to defective theoretical knowledge and practical skills in detection of oral lesions [4][5][6][7].
In order to avoid the occurrence of misdiagnosis and inappropriate management, American Academy of Oral and Maxillofacial Pathology (AAOMP) have recommended that all "abnormal tissue be submitted promptly for microscopic evaluation and analysis" underscoring the importance of histopathology as a gold standard at reaching a definitive diagnosis [8]. Histological examinations provide information not only on evidence of benign or malignant lesions but clinical behavior of lesions and prognostic information about the lesion, (which might not be achievable with clinical diagnosis) [9,10].
Referrals for biopsy of specimens in a tertiary health centre are not only from specialists (including oral and maxillofacial surgeons, periodontologists, oral medicine specialists, dermatologists and even restorative dentists), but also from general dental practitioners [11,12]. Hence, prompt disease detection and accurate diagnosis is highly essential for effective and appropriate management of pathologies. Due to paucity of documented studies that evaluates the concordance between clinical oral diagnosis and histopathological assessment in our environment, there is an overarching need to identify diagnostic concordance in our African setting, as this would provide further insight into the most efficient management of patients with oral and maxillofacial lesion [5,6]. This study therefore aims at assessing the rate of concordance between clinical and histopathological diagnosis of oral and maxillofacial lesions; and to also conduct an audit of biopsied oral lesions between the period of 2008 to 2017, which is a determinant of oral pathology service utilization; and identification of prevalent lesions in our environment. with range from 2 days to 25 years and a median duration of 1 year.

Methods
The mean duration for males (2.6 years ± 3.9) approximated but slightly higher than that of females (2.5 years ± 3.6) whereas the mean duration per decades varies significantly (p=0.002). Patients within first decade present with a mean duration of 0.9±0.9 which increased progressively to 3years ± 4.9 in the fifth decade and then tailed down to 0.2±3.6 in the tenth decade. The mean duration of presentation with presenting complaints is also very significant (p=0.0002) with mean duration of 2.7years ± 3.6 for painless swelling, 0.8 ± 0.9 for ulcerated lesions, 2.93 years ± 4.7 for painful swelling, 0.9 ± 1.8 for swelling associated with ulceration and 3.9 years ± 4 for swelling associated with purulent discharge.

Discussion
Various publications in scientific literature have reported prevalence of biopsied lesions in entirety [11,13,[16][17][18], groups [14,19,20], categories of lesions [21][22][23] and amongst anatomic locations [24,25]. Likewise, documented studies have also reported the sensitivity and specificity between clinical impressions and histological diagnosis which is the gold standard of definitive diagnosis [4,5,26]. This study aims to report the prevalence of biopsied lesions in our tertiary center and as well as the concordance/agreement rate between clinical impressions and histopathological diagnosis. The total number of biopsied lesions in our study within the period of 2008 and 2017 (10 years) was 592 cases although similar to the prevalence in a documented study [27], it is relatively low in terms of quantity to the number of biopsied lesions recorded in other studies [11,13,28,29]. This may be due to the limited number of dental facilities and few presences of private and/or general dental practices in our coverage areas. This may also be related to the overall reduced dental awareness in our populace and primitive beliefs of the Hoi-polloi to adhere to alternative practices and faith-based healing and a general anxiety or fear to any "surgical interventions". Low utilization of oral biopsies service by dental units such as pediatric dentistry and conservative dentistry could also have contributed immensely to this very low numbers.
Unlike in developed western countries where "ALL" excised tissue specimens are referred to the histopathology laboratory, the reverse is usually the case in resource limited centers [1,16].
The mean age of patients recorded in our study was 36.1years ± 18.7 which is similar to the mean age of between 34.9 years and 38 years observed in documented scientific literature [16,28,30] but lower than the mean age of 41years to 54 years observed in other studies [27,31,32]. The peak age of 3 rd and 4 th decade that was observed in this study also coincides with other studies [16,30].
Gender-wise, a slight female predilection of 54.4% was observed comparable to the reported finding of 51.4% -59% in other studies [13,27,28,30]. This may be explained in terms of higher positive attitude and oral health service utilization by women compared to male's counterpart in our environment, furthermore males tend to have a poorer health status [13]. However, there are studies with contrasting observation of male preponderances [29,33] while others observed an equal gender prevalence [31]. Difference in gender prevalence of biopsy service may not be unconnected to geographic and epidemiological demographics [16]. Furthermore, this study observed that malignant lesions and odontogenic/nonodontogenic cysts were seen more in males, finding similarly recorded in another study [17]. Similar observation of higher frequency of central(intraosseous) lesions (65%) compared to peripheral lesions in this study was also recorded in the studies of Ali et al. [26] and Fierro-Garibay [29]. This is attributed to the high prevalence of OTs and FOLs in the present study which could also be a revelation that most referred lesions for histopathology diagnosis are advanced and aggressive in nature in a resource constrained centre. Other authors have contrary findings with the observation of more peripheral and periapical pathologies of 21.5%, 16.7% and 11.4% in their studies [17,27,30].
Regarding the anatomical site, mandible with 47% was found as the most common site for biopsy, observation that was similarly reported This is at variance with Tatli et al. [6], who reported a slightly higher male concordance (0.92) compared to females (0.9). The higher female concordance in this study may be attributed to accurate account of related details and leading information of lesions by female patients. In other studies [7,42], it was opined that females tend to present earlier at the onset of the lesion, thereby making the diagnosis relatively easier.
Regarding age, highest concordance was observed within the 7 th decade and above with concordance index of 70; this was in tandem with reports in documented literature [41,43,44]. It has been proposed that reduced number of lesions in this age bracket (after exclusion of lesions that develop in children and young adults), limited number of newly developing lesions and loss of teeth at these decades might account for the high concordance observed [41].
Significantly, variable concordance was also observed with specialties, with the highest concordance index (62.5) emanating from oral periodontologists, respectively. Seifi et al. [40], in their study however recorded the highest concordance amongst oral surgeons.
The overall interrater agreement was 0.5 which was slightly higher than the previous Nigerian study by Emeka et al. [7] (0.45), but lower than 0.61 reported by Seifi et al. [5]. Intraoral sites with the highest concordance index and interrater agreement were the tongue, lower lip, buccal mucosa and mandible, respectively. This is comparable to the observations of Emeka et al. [7] and Tatli et al. [6] where the tongue, buccal mucosa and salivary glands had the highest concordance. Tatli et al. [6] attributed this to the specific characteristics of lesions in these sites. In general, concordance rates between the pathologists, surgeons and other specialties can be improved by ensuring proficiency and good interdisciplinary relationship, mostly between pathologists and surgeons (including other specialties). Accuracy of the interpretation of plain radiographs and imaging techniques; good and adequate biopsy specimen, coupled with appropriate means of conveying specimen to the laboratory, would also improve concordance [41]. Discordances should be meticulously examined by all specialties (particularly oral pathology), to improve early detection and diagnosis of diseases and biopsied lesions [6].

Competing interests
The authors declare no competing interests.