The value of 18 F-FDG PET / CT in assessment of metabolic response in esophageal cancer for prediction of histopathological response and survival after preoperative chemoradiotherapy

Aim. To evaluate the ability of hybrid 18F-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) to predict histopathological response and overall survival (OS) after preoperative neoadjuvant chemoradiotherapy (CRT) in patients with the esophageal carcinoma. Methods. 73 patients with locally advanced esophageal carcinoma were included in the study. All were treated with CRT and 34 subsequently underwent surgical resection of the esophagus. 18F-FDG PET/CT was carried out prior to (PET/ CT1) and 6 weeks after (PET/CT2) completion of the CRT. Results. PET/CT2-determined complete metabolic response (CMR) was achieved in 6 (17.6%) out of 34 operated patients, the metabolic response was incomplete (NCMR) in 28 (82.4%) patients. A histopathological complete response (CR) to CRT was discovered in 7 patients (20.6%). The median OS in operated patients was 17.1 months, 95% CI:12.9-23.3 months. In a group of 39 non-operated patients, CMR after neoadjuvant CRT was achieved in 12 patients (30.8%), while NCMR was found in 28 (82.4%). The median OS was 13.5 months in this group, 95% CI: 4.4-22.7 months. Conclusion. No statistically significant correlation was found between the 18F-FDG metabolic response after the neoadjuvant CRT and histopathological response. Presently, the contribution of 18F-FDG PET/CT as a marker of the potential result of CRT cannot be considered definite. Another study with a larger sample of patients and standardized algorithms for the examining protocols would be necessary for reaching definitive conclusions.


INTRODUCTION
Esophageal cancer ranks among the 10 most common malignancies worldwide and is associated with high mortality 1,2 -the overall survival of 5 years is only attained in 9-15% patients 3,4 .Carcinomas of the esophagus are a heterogeneous group of tumours with respect to epidemiology, etiology, and histopathology 5 .Squamous cell carcinoma (SCC) occurs most frequently in the upper two thirds of the esophagus.In the lower third of the esophagus and the esophagogastric junction, adenocarcinoma (AC) is more prevalent.
Precise pretherapeutic staging is important when choosing the best available therapy for the patient.It is crucial to be able to differentiate patients with locoregional disease from patients with systemic disease.In systemic disease, there is no curative option and the patients receive palliative treatment 6 .
After exclusion of distant metastases, selection of the therapeutic regimen depends on the T stage.Localized tumours (T1/T2) have a high likelihood of R0 resection, and primary esophagectomy represents the most frequent therapeutic procedure.In cases of locally advanced tumours (T3/T4, N+), surgery remains the mainstay of therapy, but evidence is growing that preoperative neoadjuvant chemotherapy or chemoradiotherapy (CRT) improves survival in patients with esophageal cancer [7][8][9] .The aim of neoadjuvant therapy is to cure potential micrometastatic disease with parallel preoperative tumor downsizing thereby increasing the chance of a more appropriate alternative treatment.Despite conflicting results from randomised trials, concurrent CRT followed by esophagectomy has became the standard option, with about 70% of patients receiving preoperative CRT before undergoing esophagectomy [10][11][12][13] .Of resected patients, 11-56% achieve pathologic complete response (CR), and patients who achieve pathologic CR survive longer than those who do not [14][15][16] .In a recent meta-analysis evaluating the contribution of neoadjuvant CRT for esophageal carcinoma, Gebski et al. 8 reported an absolute 2-year survival advantage of 7% with neoadjuvant chemotherapy and 13% with neoadjuvant CRT in patients with esophageal carcinomas, compared to those treated by surgery alone.The overall results of this meta-analysis also show that the benefit of neoadjuvant treatment to responders is partially negated by effects in non-responders.
The mortality after esophagectomy continues at around 5-9% (ref. 17) despite improvements in surgical technology.Furthermore, a recent European trial found that adding surgery to neoadjuvant therapy improves local tumour control but has no overall survival (OS) benefit, particularly for patients who responded to CRT (ref. 18).This finding, in combination with postoperative mortality rate from esophagectomy, also suggests that surgery may be detrimental for patients who achieved pathologic CR after preoperative CRT (ref. 19).
Taken together, converging evidence reveals that patients undergoing neoadjuvant treatment and showing an objective tumour response have a better prognosis for survival than those undergoing surgery alone.However, only 40-50% of patients respond to neoadjuvant therapy.As a consequence, the patients who do not respond to therapy may be compromised by toxic side effects and delay caused by ineffective chemotherapy or CRT, and potentially even have biologically more aggressive tumours 6 .Therefore, it is desirable to have a diagnostic test that allows noninvasive prediction of response to neoadjuvant therapy so that responders can be differentiated from non-responders.
Conventional structure-based imaging technology, such as computed tomography (CT), endoscopy and endoscopic ultrasonography (EUS), are generally considered inaccurate in predicting response to neoadjuvant treatment, in particular due to their inability to differentiate between a viable tumor and inflammation, edema or fibrosis 20,21 .The functional modality which can detect changes in tissue metabolism which usually precede structural changes, consists of 2-[fluorin-18] fluoro-2-deoxy-Dglucose (FDG) -positron emission tomography ( 18 F-FDG PET) (ref. 22,23).Current evidence suggest that this method can identify patients who achieve a pathologic response to neoadjuvant treatment prior to surgical resection, as well as being able to predict long-term survival [24][25][26] .Song et al. 27 previously showed that the pathologic response of initially highly metabolic tumour after preoperative CRT could correlate with the metabolic response.It has not, however, been unequivocally demonstrated whether a complete metabolic response (CMR), evaluated by means of 18 F-FDG PET, always reflects a histopathological complete response (CR).9][30].A number of studies have demonstrated the contribution of 18 F-FDG PET in the primary staging of esophageal carcinoma particularly in the detection of distant metastasis 28,[31][32][33][34] .The potential of this functional method in detecting response to neoadjuvant treatment in a group of patients with upper gastrointestinal malignancy was first published in 1998 (ref. 35).Subsequent studies on the ability of 18 F-FDG PET to identify histopathologi-cal responders reached conflicting conclusions [23][24][25][26][27][36][37][38][39] . Severa others found a correlation between significant reduction of 18 F-FDG accumulation in a primary lesions and a significant histopathological tumor reduction 24,25,37,39 in patients with SCC of the esophagus.
The aim of our study was to evaluate the potential of 18 F-FDG PET/CT examination carried out both prior to and after neoadjuvant chemoradiotherapy (CRT) for detecting therapeutic response in patients with esophageal carcinoma, to assess whether the level of metabolic response to chemoradiotherapy can predict a complete histopathological response and survival of these patients.

MATERIAL AND METHODS
From August 2006 to January 2010, 82 consecutive patients with biopsy-proven locally advanced squamous cell carcinoma (SCC) and adenocarcinoma (AC) of the middle and distal esophagus were included in the study.As part of staging, all patients underwent endoscopy, EUS, CT and hybrid 2-[fluorin-18] fluoro-2-deoxy-D-glucose (FDG) -positron emission tomography combined with multislice computed tomography ( 18 F-FDG PET/ CT).After these diagnostic procedures, all patients underwent neoadjuvant chemoradiotherapy (CRT).Endoscopy and 18 F-FDG PET/CT were once again carried out after completion of CRT.
Of the 82 patients who underwent 18 F-FDG PET/CT before and after CRT, 39 did not undergo esophagectomy due to patient refusal, poor general condition, disease progression or advanced age.The esophagectomy was performed in 43 patients.

F-FDG PET/CT
All patients fasted for at least 6 h and the blood glucose levels were measured.If the glucose concentration did not exceed 130 mg/dL, 400 MBq of 18 F-FDG per 70 kg of weight were administered intravenously, with the activity applied being recalculated based on the actual weight.Sixty minutes after the administration of 18 F-FDG and oral administration of a contrast medium, the PET/ CT examination using Siemens Biograph 16 HI-REZ scanner was initiated.Contrast-enhanced multislice CT scans were carried out typically from the skull base to the upper third of the thighs with arms upwards.The lung region was scanned with patients holding their breath in expiration.This was followed by caudocranial PET scanning with iterative reconstruction of the images.Transmission attenuation correction was carried out by CT.For semiquantitative analysis of each lesion showing increased 18 F-FDG uptake in the esophagus, the maximal standardized uptake value normalized to the body surface area (SUV bsa max) was computed on the most intense uptake area (graded colour-scaled parametric analysis applied in reconstructed coronal PET image) in accordance with standard formulas.

Neoadjuvant chemoradiotherapy
Neoadjuvant CRT consisted of early radiation with a linear accelerator in the area of the primary tumour and catchment nodes; total radiation was 50 Gy and was applied in 25 fractions of 2 Gy over the course of 5 weeks. 2 cycles of chemotherapy, composed of cisplatinum and fluorouracil, were applied along with radiotherapy; the third cycle of chemotherapy was applied three weeks after completion of radiotherapy.Cisplatinum was applied once a week during the period of radiotherapy in patients with a worse overall state and with SCC.Restaging (endoscopy with biopsy and PET/CT2) was carried out after completion of CRT.

F-FDG PET/CT assessment of response to neoadjuvant chemoradiotherapy
All FDG-PET/CT images were reviewed and interpreted by two experienced nuclear physicians.The presence of a normal distribution of 18 F-FDG at the site of the original pathological FDG-PET finding and even in the case of an abnormal CT finding in an identical localization was considered as a complete metabolic response (CMR).The finding of focal uptake of 18 F-FDG was considered as a non-complete metabolic response (NCMR).A diffuse increased uptake of 18 F-FDG in the esophagus in the area of radiotherapy was viewed as benign esophagitis.The visual evaluation was always accompanied by a semi-quantitative [SUV bsa max] evaluation.In cases of a negative FDG-PET finding on the pre-operative FDG-PET scans, SUV was uniformly designated as 0.5 of the baseline or background FDG uptake level 19 .

Operation and assessment of histopathological response to neoadjuvant chemoradiotherapy
After restaging, all patients underwent esophagectomy.Trans-hiatal laparoscopic exstirpation of the esophagus with gastroplasty and cervical anastomosis from laparotomy and cervical incision were carried out in patients with distal esophageal carcinoma.A lymphadenectomy and pyloromyotomy were part of the surgery.A right-sided thoracoscopic or thoracotomic exstirpation of the esophagus with passage renewal in the same fashion as described in the distal esophagus was performed for carcinomas located in the middle esophagus to an endoscopic tumour distance of 30 cm from the incisors.A classic approach (thoracotomy) was chosen for large tumours with a suspicion of infiltration of surrounding structures.
In each case the histopathological examination of the entire resection specimens were examined for degree of local tumour spread and lymph node metastases.Patients with no residual viable tumour cells in the surgical specimen (pT0N0M0) were classified as having achieved pathologic CR.Patients with macroscopic or microscopic foci of residual tumours were considered to have pathologic RD.A semi-quantitative evaluation was not used.

Statistical analysis
Survival probability analyses were performed using the Kaplan-Meier method.Survival was calculated from the beginning of neoadjuvant chemoradiotherapy (CRT) to the date of death or most recent follow-up.Statistical significance was assessed by the log-rank test.Receiveroperating-characteristic (ROC) analysis was performed to determine an optimal cut-off value of SUVmax reduction from PET/CT1 to PET/CT2 in predicting overall survival [OS].Variables age, sex, maximum SUV of primary tumour and pre-and post-CRT SUV change were used in Cox regression analyses.Data analysis was performed with the SPSS version 15.0 (SPSS, Chicago, IL).Statistical significance was defined as P<0.05.

RESULTS
Forty-three of the total 82 patients with esophageal carcinoma underwent esophagectomy.Nine died as a result of early post-operation complications and were removed from the analyses.
The median time interval from completion of the neoadjuvant CRT to the PET/CT2 was 42 days (range 21-56 days).
The groups differed significantly in distribution according to stage; the non-operated group were patients with a higher disease stage (P=0.0004,Fisher's exact test) but the difference for histopathological grade was not statistically significant (P=0.825,Fisher's exact test) (table 1).
A complete metabolic response (CMR) to neoadjuvant CRT was found in 6 patients (17.6%) in a PET/CT2 examination; a complete metabolic response (NCMR) was not achieved in 28 (82.4%)patients.
A histopathological complete response (CR) to CRT was found in 7 operated patients (20.6%), while RD was determined in 27 patients (79.4%).
According to the Cox regres sion analysis, change in SUVmax in % (RR=1.080)was a significant predictor of survival and exitus in operated patients.Decrease in SUV by less than 1% was connected with increased risk of exitus by 1.08x, 95%CI: 1.044 -1.116.
The ROC analysis yielded an optimal cut-off value of 62.4% for SUVmax reduction from PET/CT1 to PET/ CT2 in predicting OS.In the case of SUVmax2 reduction ≥ 62.4%, the overall survival could be predicted with a sensitivity, specificity, accuracy, and a positive and negative predictive value of 66.7%, 60.7%, 61.8%, 26.7% and 84.5%, respectively.The Kaplan-Meier survival analyses showed a significantly longer median overall survival in patients with SUVmax2 reduction ≥ 62.4% (log-rank test, P=0.0002, Fig. 1).
The complete metabolic response (CMR) after neoadjuvant CRT was found in 12 cases (30.8%) in the PET/ CT2 examination while the metabolic response was incomplete (NCMR) in 27 patients (69.2%).
According to Cox´s regression analysis, PET/CT2 (RR=4.377)and gender of the patient (RR=3.089)were significant predictors of OS and exitus.The results of PET/CT2 = NCMR increased the risk of exitus by 4.38x, 95% CI: 1.48 -12.92.The risk of exitus of non-operated women was 3.09 x higher than of non-operated men, 95% CI: 1.03 -9.24.
The average OS in non-operated patients with grade 1 was 31.2 months, 95% CI: 25.6 -36.8 months (the median survival cannot be estimated).The mean OS of patients with grade 2 was 13.1 months, 95% CI: 12.3-13.9months, the median survival was 13.0 months, 95% CI: 11.5 -14.6 months.The mean OS of non-operated patients with grade 3 was 8.6 months, 95% CI: 7.2 -10.1 months, the median survival was 8.2 months, 95% CI: 7.3 -9.1 months.The OS of non-operated patients was significantly different in relation to histopathological grading (log-rank test, P<0.0001, Fig. 3).

A comparison of some parameters of the groups of operated and non-operated patients
The period of OS in operated and non-operated patients in the evaluated group of 73 patients did not differ statistically significantly (log-rank test P=0.595,Kaplan-Meier analysis, Fig. 4).The average period of OS in nonoperated patients with CMR was 27.9 months, 95% CI: 19.6 -36.1 months, median survival was 22.3.The average period of OS in operated patients with CMR was 18.7 months, 95%CI: 11.8 -25.6 months, the median survival was 15.2 months, 95% CI: 15.1-15.3months.The difference in survival of both groups was not statistically significant (log-rank test, P=0.473.Fig. 5).
A significant dependence (Fisher's exact test, P=0,009) was determined between the histopathological grade of esophageal carcinoma and the metabolic response ( 18 F-FDG-PET/CT).Significantly more metabolical respondents were in the group of patients with grade 2 (46.7%) in comparison with the group of patients with grade 3 (9.1%).The results are summarized in Table 2.

DISCUSSION
We evaluated the 18 F-FDG PET/CT-determined metabolic response after neoadjuvant CRT as a marker of prediction of histopathological response and survival period in a group of 34 patients with operable esophageal carcinoma (17 SCC and 17 AC).The median time interval after completion of CRT to PET/CT2 was 42 days (range of 21-56 days).
The level of correspondence of the findings with PET/ CT2 examination (CMR/NCMR) and histopathological examination (CR/RD) was average (coefficient AC1 = 0.532).No statistically significant difference in the per- centage decrease of SUVmax2 after chemoradiotherapy (Mann -Whitney test, P=0.383) was discovered between patients with a histopathological CR and RD.Smithers et al. 40 also failed to find a correlation between 18 F-FDG PET findings after neoadjuvant chemotherapy and CRT an d a histopathological response in a group of 45 patients with AC of the esophagus.A correlation was only determined when evaluating an entire group of patients treated with neoadjuvant therapy, but not, however, when evaluating chemotherapy and CRT in dividually.The authors pointed out the possible connection between this result and the histological type of esophageal carcinoma, as studies with groups of patients with SCC of the esophagus had a significant correlation between the histopathological response and the determined metabolic ( 18 F-FDG PET) response.Brücher et al. 24 have published results on 27 patients with SCC of the esophagus.A threshold of 52% mean SUV divided the histopathological responders from nonresponders with a sensitivity of 100% and specificity of 55%.Flamen et al. 25 showed in a study of 36 patients with esophageal cancer (27 cases of SCC and 9 of adenocarcinoma) that a decrease of more than 80% in the tumour-to-liver uptake ratio 3-4 wk after completion of neoadjuvant chemoradiotherapy predicted a histopatological response with a sensitivity of 71% and a specificity of 82%.Kim et al. 19 published results of 62 patients with SCC.A complete metabolic response (reduction of maximum SUV > 80%) after completion of therapy correlated significantly with a histopathological response.However, in another study, by Cerfolio et al. 41 , with a prevalence of patients with AC of the esophagus (41 cases of adenocarcinoma and 7 of SCC) 18 F-FDG PET/CT predicted a complete histopatological response with a sensitivity of 87%, specificity of 88%, and an accuracy of 88%.The correlation between the 18 F-FDG PET response after neoadjuvant chemoradiotherapy and the histopathological response was, however, also not determined in additional studies 23,33,38 .
The authors Port et al. 42 determined the ability of 18 F-FDG PET to predict a clinical and pathological response and survival in a group of 62 patients with esophageal carcinoma (51 AC, 11 SCC).Apart from determination of a significantly improved prediction of disease-free survival in patients with a decrease in SUVmax by 50%, they also discovered that the complete absence of residual uptake of 18 F-FDG need not necessarily be connected with a complete histopathological response.Swisher et al. 23 in a group 103 patients (90 AC, 13 SCC) demonstrated that 18 F-FDG PET failed to rule out residual microscopic disease, because 18 F-FDG uptake in the tumour bed did not differ between patients with no residual viable tumour cells and patients with up to 10% viable tumour cells.A range of similar studies 23,25,36,40,43 employed the semi-quantiative Mandard's or Becker's classifications and the ir modifications 44,45 for histopathological evaluation of response to neoadjuvant therapy, where the complete histopathological response was considered either the complete absence or presence of less than 10% viable tumour cells in the resection specimens.In our study, we employed qualitative evaluation of the resected specimens where only complete absence of viable tumour cells was considered a complete histopathological response (CR) and residual disease (RD) was every histopathological finding containing viable tumour cells in any amount including smaller than 10%.It cannot thereby be fully ruled out that a less significant correlation between 18 F-FDG PET metabolic and histopathological response to CRT in our study could also have been influenced by this fact.A negative factor preventing the formulation of definitive conclusions in connection with the majority of the abovecited works could also be the limited number of patients in the groups.
Krause et al. 6 in their review work evaluating the importance of 18 F-FDG PET and 18 F-FDG PET/CT examinations for evaluation of response to treatment of esophageal carcinoma state that most studies assessing a late response to neoadjuvant therapy for esophageal cancer (3-6 weeks after completion of neoadjuvant therapy) have shown that 18 F-FDG PET signal after neoadjuvant therapy correlates with histopathological response and long term prognosis.However, the main drawback of late assessment is that it does not allow therapy modification for patients not responding to it.One might speculate whether patients benefit from a change in therapy after a late response assessment.The authors emphasize that clinical trials are necessary to answer the question of whether further neoadjuvant or adjuvant chemotherapy or chemoradiation for 18 F-FDG PET non-responders improves clinical outcome.Therefore, an early assessment of response to therapy by 18 F-FDG PET has been proposed as a surrogate marker for predicting response and potentially allowing therapy modification, although is not yet established in clinical routine practice.Weber et al. 36 reported a histopathological response prediction with a sensitivity of 93% and a specificity of 95% in 40 patients with AC of the esophagogastric junction undergoing neoadjuvant chemotherapy. 18F-FDG PET images were performed pre-treatment and 14 days after commencement of neoadjuvant treatment in contrast to the preoperative 18 F-FDG PET performed some weeks after the conclusion of treatment.Thus the timing of the scan may be critical, with earlier scans during therapy providing a clinical guide in patients with esophageal cancer.Similar results with early (2 weeks) assessment of pathologic response to neoadjuvant chemotherapy of locally advanced adenocarcinoma of the esophagus and esophagogastric junction has been reported in the prospective unicenter MUNICON study 46 .
According to the results of Cox's regression analysis the parameter of change in SUVmax in % (RR=1.080)and the histopathological grade of the tumour (log-rank test, P<0.0001) were significant predictors of survival in our group of 34 operated patients.Tumour grading was also a significant predictor of the overall survival in the group of 39 non-operated patients (log-rank test, P<0.0001).
The survival of operated (34) and non-operated (39) patients in our set of 73 patients did not differ signifi-Long rank test P=0.473cantly.Despite the fact that the median of OS patients with CMR in both groups was relatively different on average, the difference is not significant in light of the small number of patients (log-rank test, P=0,473).

CONCLUSION
In conclusion, it can be stated that in our group of a mixed population of operated patients with SCC and AC of the esophagus we did not find, in contrast to several similar studies, a robust correlation between 18 F-FDG metabolic response after neoadjuvant CRT and histopathological response, and the contribution 18 F-FDG PET/CT as a marker of the potential result of CRT cannot at present be considered unequivocal.In light of the non-homogenous character of the examining protocols, the varied regimens of neoadjuvant chemotherapy and finally the small number of treated patients in the compared groups, further studies with a larger number of patients and with standardized of protocols will be necessary in order to reach definitive conclusions.In the group of operated patients, 18 F-FDG PET/CT examination predicted a significantly longer period of survival in patients with a reduction of SUVmax ≥ 62.4%.This result is in accordance with the majority of published works.Surprisingly, a statistically significant difference in overall survival between operated and non-operated patients was not demonstrated.

Fig. 3 .Fig. 4 .
Fig. 3.The Kaplan-Meier survival probability analysis shows a statistically significant difference in OS in the group of nonoperated patients (N = 39) in relation to histological grading (log-rank test, P<0.0001).

Fig. 2 .Fig. 5 .
Fig. 2. The Kaplan-Meier survival probability analysis shows a significant difference in OS in the group of operated patients (N = 34) in relation to histopathological grading (log-rank test, P<0.0001).

Table 1 .
Baseline demographic and clinical characteristics of the entire group of patients with esophageal cancer, groups of operated and non-operated patients (N = 73).