Estimated projection of oral squamous cell carcinoma annual incidence from twenty years registry data: a retrospective cross-sectional study in Indonesia

Background The incidence of oral squamous cell carcinoma (OSCC) has not been well documented in Indonesia. Thus, we aimed to analyze trends and clinicopathological profiles of OSCC cases in Indonesia, focusing on differences between age and sex groups. Methods A cross-sectional study was conducted in Indonesia’s main referral hospital, analyzing 1,093 registered OSCC cases from 2001 to 2020. Trend analysis was performed using Joinpoint regression analysis to determine the annual percentage change (APC) for overall cases and each case group based on age, sex, and anatomical subsites. APC significance was assessed using a Monte Carlo permutation test. The projection of case numbers for the following 5 years (2021–2025) was estimated using linear/non-linear regression analysis and presented as a mathematical function. The significance of the trend slope was measured using an ANOVA test. Demographic and clinicopathological characteristics of OSCC were analyzed according to age and sex, and their comparative analysis was assessed using Chi-square and its alternatives. Results The incidence of OSCC in female patients and in the tongue and buccal mucosa showed a positive trend (APC 2.06%; 3.48%; 8.62%, respectively). Moreover, the incidence of OSCC overall, and in women with OSCC, is projected to increase significantly in the next 5 years following the quadratic model. The mean age of patients was 51.09 ± 14.36 years, with male patients being younger than female patients. The male-to-female ratio was 1.15, and 36.5% of these patients were categorized as young (≤45 years old). The tongue was the predominantly affected site. Prominent pathologic characteristics included well-differentiation, keratinization, and grade I of Bryne’s (1992) cellular differentiation stage. Most patients presented with advanced staging, lymphovascular invasion, and uninvaded margins. Tumor sites and staging varied according to age, while age and tumor sites differed between sexes. Conclusion The rising incidence trends of OSCC among Indonesian patients, both in the past and projected future, are concerning and warrant attention. Further research into risk factors should be conducted as preventive measures.


Interpretation:
The fitted and residual data of overall patients show a correlation near zero, indicating stationarity, independence, and absence of systematic patterns.This suggests that the model effectively captures underlying data patterns and produces accurate forecasts.

Interpretation:
• The normality test for the residuals yielded a p-value of 0.851, suggesting a close alignment with the red line ("zero line" or "residuals mean line" or "identity line")and indicating a normal distribution.• The R-square value for the Quadratic model is 20.1%, indicating a reasonable fit of the regression model to the data.• The plot of fitted values versus residuals shows consistent variance across all fitted values, suggesting the assumption of homoscedasticity is likely satisfied.

Interpretation:
The fitted and residual data of young patients show a correlation near zero, indicating stationarity, independence, and absence of systematic patterns.This suggests that the model effectively captures underlying data patterns and produces accurate forecasts.

Interpretation:
The fitted and residual data of old patients show a correlation near zero, indicating stationarity, independence, and absence of systematic patterns.This suggests that the model effectively captures underlying data patterns and produces accurate forecasts.

Interpretation:
The fitted and residual data of male patients show a correlation near zero, indicating stationarity, independence, and absence of systematic patterns.This suggests that the model effectively captures underlying data patterns and produces accurate forecasts.

Interpretation:
The fitted and residual data of female patients show a correlation near zero, indicating stationarity, independence, and absence of systematic patterns.This suggests that the model effectively captures underlying data patterns and produces accurate forecasts.

4). The results of
2).Residual analysis, goodness-of-fit, and homoscedasticity for the Quadratic modelInterpretation:• The normality test for the residuals yielded a p-value of 0.684, suggesting a close alignment with the red line ("zero line" or "residuals mean line" or "identity line") and indicating a normal distribution.• The R-square value for the Quadratic model is 24.0%, indicating a reasonable fit of the regression model to the data.• The plot of fitted values versus residuals shows consistent variance across all fitted values, suggesting the assumption of homoscedasticity is likely satisfied.
analysis, goodness-of-fit, and homoscedasticity for the Quadratic model Interpretation: • The normality test for the residuals yielded a p-value of 0.291, suggesting a close alignment with the red line ("zero line" or "residuals mean line" or "identity line") and indicating a normal distribution.• The R-square value for the Quadratic model is 14.4%, indicating a reasonable fit of the regression model to the data.• The plot of fitted values versus residuals shows consistent variance across all fitted values, suggesting the assumption of homoscedasticity is likely satisfied.
analysis, goodness-of-fit, and homoscedasticity for the Quadratic model Interpretation: • The normality test for the residuals yielded a p-value of 0.124, suggesting a close alignment with the red line ("zero line" or "residuals mean line" or "identity line")and indicating a normal distribution.• The R-square value for the Quadratic model is 34.1%, indicating a reasonable fit of the regression model to the data.• The plot of fitted values versus residuals shows consistent variance across all fitted values, suggesting the assumption of homoscedasticity is likely satisfied.

Regression Model for OSCC Cases in Old Patients (>45 years old) 1). Curve options
Yt = 28.91 + 0.552t 17.8266 6.1608 51.1830 Quadratic model* Yt = 34.48-0.97t + 0.0723t 2 16.2040 5.5500 46.5894Growth curve model Yt = 29.1201× (1.01457t) 16.9183 5.9632 51.3219 *The best-fitted model is the one that has the lowest value for three parameters (MAPE, MAD, and MSD), or at least for two parameters out of three, or at least having the lowest value for MAPE.

Regression Model for OSCC Cases in Male Patients 1). Curve options
Yt = 24.71+ 0.432t 19.1804 5.6575 55.5727 Quadratic model* Yt = 28.98 -0.73t + 0.0554t 2 18.8513 5.4835 52.8764Growth curve model Yt = 24.798× (1.01273 t ) 17.9393 5.4787 56.2548 *The best-fitted model is the one that has the lowest value for three parameters (MAPE, MAD, and MSD), or at least for two parameters out of three, or at least having the lowest value for MAPE.

.16 -0.729t + 0.0620t 2 15.4506 3.9259 27.6543
*The best-fitted model is the one that has the lowest value for three parameters (MAPE, MAD, and MSD), or at least for two parameters out of three, or at least having the lowest value for MAPE.