Page 45 - ARNM-1-2

Advances in Radiotherapy
& Nuclear Medicine SUVmax relating to patient and tumor factors

iterative algorithm (OSEM: accelerated ordered subsets According to the World Health Organization
expectation maximization implementations and iterative criteria, tumor differentiation is classified into three
reconstruction) and reformatted into transaxial, coronal, histological grades (well, moderate, or poor). In our
and sagittal views. Fusion of PET and CT images was study, 24 patients had well-differentiated tumors, 53 had
achieved using the specialized Siemen’s fusion software. moderately-differentiated tumors, and 34 had poorly-
differentiated tumors. The actual distribution with respect
2.3. Outcome parameters to various tumor sites, histology, and grades is presented in
Demographic data (age and gender), clinical data (tumor Tables 1 and 2. The frequency distributions of SUV for
max
site and stage), and pathological data (histopathology and age, gender, sites, histology, and tumor grade are presented
grades) were recorded, and the correlation of SUV to in Tables 3 and 4.
max
these parameters was analyzed. Based on Spearman’s correlation test, a positive
correlation was observed between histology and the SUV ,
2.4. Statistical analysis max
with P-value of 0.01. Likewise, the chi-square test revealed
Our data were analyzed using the IBM SPSS Version a significant positive correlation between histology and
20 (the Statistical Package for the Social Sciences, SUV with P-value of 0.024 (Table 5). However, no such
max
International Business Administration Corporation, USA) association was detected between SUV and other factors
max
software. Categorical variables were compared using such as age, gender, system, grade, and stage.
Pearson’s Chi-square test, and correlations were assessed The regression model (Equation I) was constructed
using Spearman’s correlation test. The confidence intervals using grade, age (years), and system.
were set at 95%, and P < 0.05 was considered statistically
significant. The validity of the regression equation was SUV =12.728 + (0.33 × Age [year] + (0.158 × Grade)
max
verified, and non-correlated independent variables were − (0.515 × System) (I)
selected for the forward stepwise regression model. This model shows a better predictive value compared
to other predictors (P < 0.05) (Table 6). In the probability-
3. Results probability plot, the predicted outcome follows the diagonal
As illustrated in Figure 1A and B, a total of 117 patients line, which fulfills the criteria for normal distribution
(63 males and 54 females) with a median age of 61 years (Figure 2A), and the predicted values concentrate in a single
(range: 21 – 90) were included in the study. The majority of area similar to the original values (Figure 2B). This model
the patients were in the seventh decade of age (33%) followed was constructed based on the bragging methodology, which
by the sixth decade (28%). Out of the 117 patients, 29% is appropriate for accurate prediction in our study sample.
had metastatic cancers (Figure 1C). Histopathologically,
the tumors were classified into three types: squamous 4. Discussion
cell carcinoma in 71% patients, adenocarcinoma in 24% Several studies have investigated the relationship between
patients, and small cell cancer in 5% patients. the SUVmax of FDG PET/CT and demographic, clinical,

A B

C

Figure 1. (A) Age, (B) gender, and (C) cancer stage distribution; n = 117.

Volume 1 Issue 2 (2024) 3 https://doi.org/10.36922/arnm.2032