Additional Value of 2-[18F]FDG PET/CT Comparing to MRI in Treatment Approach of Anal Cancer Patients

Accurate staging and treatment planning are imperative for precise management in Anal Cancer (ACa) patients. We aimed to evaluate the additive and prognostic value of pre-treatment 2-[18F]fluoro-2-deoxy-d-glucose positron emission tomography/computed tomography (2-[18F]FDG PET/CT) in the staging and management of ACa compared to magnetic resonance imaging (MRI). This retrospective study was conducted on 54 patients. Pre-treatment 2-[18F]FDG PET/CT studies and MRI reports were compared considering the primary tumor, pelvic lymph nodes, and metastatic lesions. The impact of 2-[18F]FDG PET/CT in the management and its prognostic value, using maximum standardized uptake value (SUVmax), were assessed. Discordant findings were found in 46.3% of patients (5 in T; 1 in T and N; 18 in N; and 1 in M stage). 2-[18F]FDG PET/CT resulted in up-staging in 9.26% and down-staging in 3.7% of patients. Perirectal lymph nodes were metabolically inactive in 12.9% of patients. Moreover, 2-[18F]FDG PET/CT resulted in management change in 24.1% of patients. Finally, SUVmax provided no prognostic value. 2-[18F]FDG PET/CT altered staging and management in a sizable number of patients in this study, and supports a need for a change in guidelines for it to be used as a routine complementary test in the initial management of ACa.


Introduction
Anal Carcinoma (ACa) is a rare cancer, consisting of 2.5% of the colorectal malignancies and approximately 0.5% of all cancers. The cancer-related death rate is only 0.2% per year [1,2]. However, with an extended field-of-view, full-ring high-resolution LSO PET component and a 128-slice spiral CT component. All patients had undergone simultaneous diagnostic CT scans, which were contrast enhanced in 74.1% of the patients. 2-[ 18 F]FDG PET/CT studies were thoroughly re-evaluated to obtain information regarding the size and maximum standardized uptake value (SUVmax) of the primary tumor, involved lymph nodes and metastatic lesions. The initial 2-[ 18 F]FDG PET/CT reports were also reviewed and the final interpretation was based on consensus. An advanced PET/CT software (AW-4.4 and 4.6; GE Medical Systems, Milwaukee, WI, UAS) permitting PET, CT and fusion PET/CT data to be viewed simultaneously, was used for reading. According to the physiologic 2-[ 18 F]FDG activity in the anal region, only remarkably increased tracer uptake (interpreted as pathologic by board-certified nuclear medicine specialists) was considered tumoral and used for determination of the size and SUVmax (determined on PET and fused images). Moreover, CT images were utilized in categorizing the lymph nodes with borderline activity into malignant or benign findings. In addition, invasion to adjacent organs was more accurately defined using concurrent CT. Bidimensional axial and coronal metabolic diameters of the PET lesions were measured using a cut-off of more than 50% tracer intensity. All metabolic diameters were correlated with the morphological diameter on CT for improving the accuracy and reliability of the measurements. However, for the data analysis, the longest dimension of the primary tumor and the short axial dimension of the 2-[ 18 F]FDG avid lymph nodes was correlated with the corresponding dimension on MRI. Larger lymph nodes with mild 2-[ 18 F]FDG uptake, some with fatty hilum and/or unsuspicious pattern on CT, were categorized as reactive or inflammatory. On the other hand, the subcentimetric spherical lymph nodes showing a lower degree of 2-[ 18 F]FDG uptake were categorized as metastatic. The final interpretation of the lymph nodes was mainly based on the experienced nuclear medicine physician's interpretation; however, we did not use any SUV cut-off for classification of the lymph nodes as benign or malignant. In case of discrepancy between the primary report and the repeating review, a senior non-primary interpreter was involved and the final result recorded based on a consensus.
Lesions were evaluated in three groups: primary tumor, lymph nodes, and metastatic lesions. Regarding the well-established fact that MRI is a superior modality in precisely defining the primary tumor 8 , the data regarding lymph nodes and metastatic lesions are mainly compared and discussed.
Lymph nodes were evaluated in four groups, and total 5 regions, according to their impact on radiation field or dose: 1 anorectal, perirectal, and paravertebral; 2 and 3 internal and external iliac (right and left); and 4 and 5 inguinal (right and left). Eventually, the Tumor, Nodes, and Metastases (TNM) classification (version 8) and stage were defined according to the TNM classification of malignant tumors.

MRI
The data were exploited from available MRI reports. MRI standard protocol in our institution includes coronal T2-weighted short-TI inversion recovery of the pelvis, multiplanar noncontrast small field of view T2-weighted turbo spin echo sequences and diffusion weighted imaging with b-values of 50 and 800, axial non-contrast T1-weighted turbo spin echo and gadolinium-enhanced T1-weighted turbo spin echo with fat saturation sequences. The contrast-enhanced sequence was used for better evaluation of lesion enhancement, to highlight the anatomical delineation of adjacent structures and also to emphasize tumor relation to the sphincters.
In 13% (7/54) of patients, data regarding MRI protocol was not available. However, 38.9% (21/54) of patients had undergone conventional and Diffusion Weighted MRI based on the above-mentioned MRI imaging protocol and 48% (26/54) had only conventional MRI with different protocols. The average interval between MRI and FDG PET/CT studies was 12.5 ± 13.9 days (median: 9, range: 0-56 days, interquartile range: 1-18 days). Except in two patients, MRI imaging was conducted before and after contrast medium administration. Lesions were evaluated according to the aforementioned categories. Afterward, the TNM classification and stages were defined.

Therapeutic Approach
The therapy information was recorded using the database and patient notes. Regarding the therapeutic approach, 85.2% (46/54) of the patients had been subjected to chemotherapy with 5-FU/Xeloda and Mitomycin-C concomitant with radiation-therapy, 1.9% (1/54) received only chemotherapy, 1.9% (1/54) underwent palliative chemoradiotherapy, and 11.2% (6/54) received only radiation-therapy. Therapeutic approaches were re-evaluated separately considering MRI and 2-[ 18 F]FDG PET/CT staging, and were documented. The practiced therapy was considered as the standard approach.

Standard of the Truth
Because of the retrospective nature of the study and also due to ethical consideration, histopathologic evaluation of detected lesions was not available. In case of discordant findings between MRI and 2-[ 18 F]FDG PET/CT, information from other imaging modalities such as ultrasonography, previously acquired CT or MRI, as well as subsequent imaging results during the follow-up, were employed to determine the false or true findings, regarding lymph node and distant metastases. The judgment was based on a consensus. MRI was considered the gold-standard for T-staging.

Follow-Up
All patients were followed using the database and asking their treating physicians for imaging reports and patient notes with a mean of 41.5 ± 29.3 (range: 2-114) and median of 37 months. The follow-up information was used for defining overall survival (OS) and disease-free survival (DFS). The end-point for OS was death and for DFS were recurrence or metastasis. Approximately thirteen percent of the patients (12.9%; 7/54) did not respond to treatment, requiring further therapeutic or palliative procedures (died: 2; salvage surgery: 2; additional radiation therapy: 3).

Statistical Analysis
Numerical data are presented as means ± SD, as well as median with an interquartile range (IQR) is provided for all variables with a nonparametric distribution. Only SUVmax of the primary tumor showed a normal data distribution using the Kolmogorov Simonov test, and the results were compared using the t-test. Other variables were analyzed using the nonparametric tests. Disease-free survival (DFS) was correlated with SUVmax using cox-regression test. A p-value of less than 0.05 was considered significant. Also, for evaluation of agreement between 2-[ 18 F]FDG PET/CT and MRI for T-stage, Cohen's kappa value was used (a value more than 0.8 suggests almost perfect agreement). The KaplanMeier analysis used for plotting the DFS. Chi-square or Fisher's exact test was applied to categorical data when appropriate. Statistical analysis was conducted with dedicated software (SPSS 23.0; IBM Corp., Armonk, NY, USA).

Results
Overall, 54 patients (81.5% female and 18.5% male with the mean age of 61.0 ± 10.6) were included for analysis of this study ( Figure 1). J. Clin. Med. 2020, 9, x FOR PEER REVIEW 5 of 15

Results
Overall, 54 patients (81.5% female and 18.5% male with the mean age of 61.0 ± 10.6) were included for analysis of this study ( Figure 1).    F]FDG PET/CT. Metastatic lesions were detected in 2/54 cases (3.7%) involving liver (one patient), and bone marrow with liver, lung and retroperitoneal lymph nodes (one patient). Details regarding TNM classification are depicted in Table 2.

Imaging Results
The mean and median size of the primary tumor was 45.9 ± 29.6 and 40.0 (IQR: 22-61) mm (metabolic diameter), respectively, on 2-[ 18 F]FDG PET/CT and 43.0 ± 26.2 and 35.0 (IQR: 25-62) mm, respectively, on MRI. The mean SUVmax was 12.1 ± 6.4 for primary lesions and 7.0 ± 5.3 for the most prominent involved lymph nodes for each patient. Details are provided in Table 3. The mean SUVmax of the primary tumor was higher in patients with higher TNM stages, although this relationship was not statistically significant (p = 0.06). Overalls, SUVmax of the primary tumor was lower in patients with complete response to treatment (11.4 ± 6.6 vs. 14.6 ± 5.1); however, it did not show statistical significance (p = 0.121). Likewise, SUVmax of the primary tumor was higher in the group with no response to treatment (16.0 ± 3.9 vs. 11.5 ± 6.5; p = 0.081) (2 with stage IV, 4 with IIIC and 1 with IIIA). No relationship was found between SUVmax and incidence of local recurrence (p = 0.84); however, the number of patients was only two. There was a significant inverse relation between lymph node as well as metabolic lymph node involvement and complete response to treatment (p = 0.006).  * The primary lesion had already been excised in some patients before performing the imaging.

Discordant Findings
In 46.3% (25/54) of the patients, discordant findings were found between MRI and 2-[ 18 F]FDG PET/CT results (5 in T; 1 in T and N; 18 in N; and 1 in M stage). However, these discordant results led to alterations in staging or management in only some.

Discordant Findings
In 46.3% (25/54) of the patients, discordant findings were found between MRI and 2-[ 18 F]FDG PET/CT results (5 in T; 1 in T and N; 18 in N; and 1 in M stage). However, these discordant results led to alterations in staging or management in only some.
On the other hand, 2-[ 18 F]FDG PET failed to accurately delineate stage in 11.2% (6/54) of patients leading to erroneous up-staging in 1.9% (1/54) and down-staging in 9.3% (5/54). Of note, excluding the incompatible findings regarding the tumor size, which is best delineated on MRI, only two cases (3.7%) were down-staged by 2-[ 18 F]FDG PET/CT. This finding was attributed to perirectal lymph nodes detected only on MRI. The information regarding discordant findings of the stage is listed in Table 5.

Survival
The OS was 96.30% for all stages within 41.5 ± 29.3 months of follow-up. About 77.8% of the patients were disease-free up to the end of the study (Figure 3). No significant correlation was demonstrated between DFS and SUVmax of the primary tumor (p-value = 0.127), SUVmax of the most prominent lymph node (p = 0.478), inguinal lymph nodes (p-value = 0.552) or iliac lymph nodes (p-value = 0.250).
The OS was 96.30% for all stages within 41.5 ± 29.3 months of follow-up. About 77.8% of the patients were disease-free up to the end of the study (Figure 3). No significant correlation was demonstrated between DFS and SUVmax of the primary tumor (p-value = 0.127), SUVmax of the most prominent lymph node (p = 0.478), inguinal lymph nodes (p-value = 0.552) or iliac lymph nodes (pvalue = 0.250).

Discussion
Anal carcinoma is a rare cancer with a five-year survival of 68.3% [2]. The precise assessment of tumor burden is of substantial importance in the management of ACa. MRI is widely accepted and employed in the initial staging of ACa. It plays a significant role in the evaluation of response to treatment, recurrence, and survival [8]. Despite its substantial advantages, it employs an expensive and time-consuming procedure with administration of contrast medium. 2-[ 18 F]FDG PET/CT is widely used in oncology and, although not routinely recommended by guidelines, however, based on the results of this study, it may provide some additional information when performed for initial staging of ACa patients.

Additive Value of 2-[ 18 F]FDG PET/CT in Staging
The role of 2-[ 18 F]FDG PET/CT seems more prominent in the evaluation of nodal and distant metastases. Comparing with conventional imaging, in a meta-analysis, Jones et al. showed that 2-[ 18 F]FDG PET/CT up-stages the nodal disease in 21% (95% CI 13-30) and down-stages in 17% (95% CI 11-23) of patients [32]. Respecting the overall stage, Mistrangelo et al. and Winton et al. reported the change in 62.5% and 23% of patients, respectively, in comparison to conventional imaging [7,16]. Moreover, in comparison to pelvic MRI plus thoracic and abdominal CT, Wells et al. and Bhuva et al. reported the overall change in stage in 47% and 42% of patients, respectively [12,14]. Similar to our study, these changes mainly attributed to nodal involvement [7,12,14,16]. We detected true alteration of the stage in 13.0% (7/54) of cases, (an increase in 9.3% and a decrease in 3.7%). In addition, 2-

Discussion
Anal carcinoma is a rare cancer with a five-year survival of 68.3% [2]. The precise assessment of tumor burden is of substantial importance in the management of ACa. MRI is widely accepted and employed in the initial staging of ACa. It plays a significant role in the evaluation of response to treatment, recurrence, and survival [8]. Despite its substantial advantages, it employs an expensive and time-consuming procedure with administration of contrast medium. 2-[ 18 F]FDG PET/CT is widely used in oncology and, although not routinely recommended by guidelines, however, based on the results of this study, it may provide some additional information when performed for initial staging of ACa patients.

Additive Value of 2-[ 18 F]FDG PET/CT in Staging
The role of 2-[ 18 F]FDG PET/CT seems more prominent in the evaluation of nodal and distant metastases. Comparing with conventional imaging, in a meta-analysis, Jones et al. showed that 2-[ 18 F]FDG PET/CT up-stages the nodal disease in 21% (95% CI 13-30) and down-stages in 17% (95% CI 11-23) of patients [32]. Respecting the overall stage, Mistrangelo et al. and Winton et al. reported the change in 62.5% and 23% of patients, respectively, in comparison to conventional imaging [7,16]. Moreover, in comparison to pelvic MRI plus thoracic and abdominal CT, Wells et al. and Bhuva et al. reported the overall change in stage in 47% and 42% of patients, respectively [12,14]. Similar to our study, these changes mainly attributed to nodal involvement [7,12,14,16]. We detected true alteration of the stage in 13.0% (7/54) of cases, (an increase in 9.3% and a decrease in 3.7%). In addition, 2-[ 18 F]FDG PET/CT erroneously up-staged one patient (1.9%) and down-staged five (9.3%). In the former, 2-[ 18 F]FDG PET/CT over-estimated the size of the primary tumor and resulted in an increase in T stage. This could be due to 2-[ 18 F]FDG uptake in surrounding inflammatory tissue. However, it did not impact the management. In two patients down-staged by 2-[ 18 F]FDG PET/CT, there was inadequate visualization of invasion to surrounding tissues, and in one case it under-estimated the tumor size. Since the management of T3 and T4 stages is similar, the therapeutic plan did not change. Though target volume delineation with both 2-[ 18 F]FDG PET/CT and MRI has demonstrated comparable results based on the findings of our study and previous investigations [33], it is well-established that MRI is a more accurate imaging modality in determination of T stage [8]. According to our results, stage alteration was seen in fewer patients, yet within the reported range. This is predominantly attributed to the recent change in TNM classification, in which lymph node involvement of any region is considered N1, influencing the stage equally. Whereas, in the previous TNM staging, mesorectal (N1) and unilateral (N2) or bilateral (N3) inguinal or iliac lymph nodes involvement pertained to different stages.
In the other two patients, the falsely down-staging by 2-[ 18 F]FDG PET/CT was attributed to perirectal lymph nodes. Of note, there were small mesorectal lymph nodes in seven patients not detected by 2-[ 18 F]FDG PET/CT. This is in part owing to the small size of the lymph nodes, below the spatial resolution threshold of detection by PET (median size = 6 mm). Moreover, intense 2-[ 18 F]FDG uptake in the primary tumor, as well as variable physiologic uptake in the gastrointestinal system, can conceal these small lymph nodes. The inferior detection rate of 2-[ 18 F]FDG PET/CT in perirectal lymph nodes has been reported [5,14,16]. Generally, mesorectal lymph nodes are included in the radiation field of the primary tumor. Therefore, the presence of these small lymph nodes does not change management, unless they appear in the T1 stage, which amends the stage from I to IIIA, as well as radiation dose from 45 to 54-59 Gy [34]. In the current study, only two patients had an absence of sub-centimetric mesorectal lymph nodes on 2-[ 18 [7]. They reported only 9.7% false positive and 4.9% false negative findings for inguinal lymph node involvement on 2-[ 18 F]FDG PET/CT [7]. Moreover, in a recent meta-analysis, the sensitivity and specificity of 2-[ 18 F]FDG PET/CT for the detection of inguinal lymph nodes was 93% and 76%, respectively [24]. ACa is rare cancer, making it challenging to conduct a study on a large population; additionally, the histopathological examination of abnormal findings detected on imaging is not a routine procedure. Therefore, the data in this regard is scarce and divergent [5,7,16,35]. Although an imperfect measure, we used clinical and follow-up data to determine false versus true findings. Indeed, the major limitation of 2-[ 18 F]FDG PET/CT-based radiation planning is its false positive findings. Further multidisciplinary multicenter studies are required to establish an accurate diagnostic performance.

Additive Value of 2-[ 18 F]FDG PET/CT in Treatment Approach
The  [7,15]. Also, in another study by Wells et al., not only management modification was reported in 37% of the initial therapy planning, but 2-[ 18 F]FDG PET/CT also changed the approaches in 17.0% of patients in post-treatment evaluation procedures [12]. Additionally, a recent meta-analysis demonstrated modification in treatment in 12.5-59.3% of patients, mainly attributed to the radiation dose or field [24]. In line with reported data, our study showed that 2-[ 18 F]FDG PET/CT enhanced management in 24.1% of patients by finding a distant liver metastasis in 1.9% and additional regions of involved lymph nodes in 18.5%, as well as by reducing dispensable radiation to inguinal regions in 3.7% (Figures 4 and 5). However, the slightly higher rate could be due to a lack of histopathology and a higher rate of false positive lymph nodes for which we could neither confirm nor rule out their involvement.       [15,16,37]. In our study, 2-[ 18 F]FDG PET/CT found one (1.9%) case with distant metastases (liver) that led to a significant adjustment of therapy (a change to palliative treatment). The lesion was primarily missed on conventional imaging modalities. Also, 2-[ 18 F]FDG PET/CT revealed additional lung metastases in a patient with known liver involvement apparent on both MRI and 2-[ 18 F]FDG PET/CT. Insofar as MRI routinely done for evaluation of the pelvic region, whole-body MRI would be of more relevance to be compared with 2-[ 18 F]FDG PET/CT in the detection of distant metastasis. However, performing whole body MRI seems impractical for all patients. Although, MRI is the optimal imaging modality for T-Staging of the anal cancer; however, if we do standard staging procedures (e.g., pelvic MRI plus thorax and abdomen CT) to pelvic MRI plus whole body 2-[ 18 F]FDG PET-CT, the latter seems to provide more accurate staging in ACa patients. Nevertheless, it was not the main objective of this study and should be evaluated in future researches.

Value of 2-[ 18 F]FDG PET/CT in Prognosis
The prognostic value of the intensity of 2-[ 18 F]FDG uptake (SUVmax) in the prediction of prognosis has been demonstrated in other malignancies [38,39]. Regarding ACa, the findings are inconsistent. In one report, the higher pre-treatment SUVmax was associated with poorer DFS [29]. On the other hand, several other studies failed to demonstrate a relation between SUVmax of the tumor and survival parameters [26,27,30,40]. In line with their findings, our results showed that higher tumor SUVmax on the pre-treatment 2-[18F]FDG PET/CT did not predict worse DFS (p = 0.127). However, SUVmax of the tumor on the post-treatment scan seems more accurate in the evaluation of outcome [26,41,42].
In recent studies, other metabolic parameters such as metabolic tumor volume (MTV) or total lesion glycolysis (TLG) have also been evaluated, which seemingly are more predictive than conventional metabolic parameters. Gauthé et al. revealed that MTV of the pre-treatment 2-[18F]FDG PET/CT is significantly correlated with OS [43]. Jones et al. claimed that the only parameter predicting any recurrence was MTV of 41% maximum SUV on the pre-treatment 2-[18F]FDG PET/CT [40]. Also, Leccisotti et al. documented that higher pre-treatment MTV and TLG are correlated with poorer different survival parameters [27].
It is well-documented that inguinal lymph node involvement, especially bilateral, worsens prognosis in ACa [31,44]. Similarly, in our patients, metabolic lymph node involvement was correlated with recurrence and no response to treatment (p = 0.006). However, no statistically significant relationship was demonstrated between SUVmax of the lymph nodes and DFS.

Limitations
The major limitation of this study is the lack of histopathology confirmation of detected lymph nodes. However, we attempted to bring all the information together and reached a consensus to minimize errors. Another limitation is the rather low number of patients, although we looked for all available data and enrolled referred patients for approximately 10 years, reflecting the low prevalence of ACa. Moreover, the number of patients that died or had only local recurrence was too small (2 patients in each group); hence, we could not assess their correlation with SUVmax. Moreover, the prognostic value of SUVmax may be compromised due to the limited number of patients. Also, the short duration of follow-up in some patients could be problematic since some of the recurrences could have occurred after the termination of our study period for follow-up. Another limitation is the use of two different PET/CT scanners, which may adversely affect the quantitative measurements. To reduce the impact, we reviewed and re-analyzed all PET/CT scans. In addition, in our study, we used MRI reports, some have been performed in other institutions, which may cause bias. However, we believe that this had a minimal effect since MRI images are reviewed by a radio-oncologists before initiation of therapy and these reports are used to determine the therapeutic approaches in the clinical practice. Finally, an interval of maximum eight weeks between two imaging modalities may affect the results for drawing an accurate statement; however, most of the included patients (49/54, 91%) were examined by both 2-[ 18 F]FDG PET/CT and MRI within four weeks.

Conclusions
2-[ 18 F]FDG PET/CT revealed influential information for more accurate staging in 12.9% of patients, and more importantly, led to management change in 24.1%, mainly in the determining of the radiation field or dose. MRI was superior in the detection of anorectal lymph nodes, but the presence of these lymph nodes is of little value in directing management. No statistically remarkable prognostic advantage was demonstrated for SUVmax of the lesions on the pre-treatment 2-[ 18 F]FDG PET/CT. 2-[ 18 F]FDG PET/CT is not expected to replace MRI. However, the invaluable potential role of 2-[ 18 F]FDG PET/CT in the management of anal carcinoma may advocate for its routine use, along with pelvic MRI, in the clinical practice.