2-Fluoro-2-deoxy-D-glucose positron emission tomography imaging is predictive of pathologic response and survival after preoperative chemoradiation in patients with esophageal carcinoma
Abstract
BACKGROUND
The current study was performed to assess the value of 2-fluoro-2-deoxy-d-glucose positron emission tomography (FDG-PET) in predicting the pathologic response and survival of patients with esophageal carcinoma treated with preoperative chemoradiation (CRT) and tumor resection. Preliminary reports suggest that FDG-PET may be predictive of the response of esophageal carcinoma patients to preoperative CRT.
METHODS
Eighty-three patients with resectable esophageal carcinoma who underwent preoperative CRT and FDG-PET and tumor resection were evaluated for pathologic response to CRT, percent residual tumor, and survival.
RESULTS
The majority of patients in the current study were men (74 of 83 patients; 89%). Most tumors were adenocarcinomas (73 of 83 tumors; 88%) and clinical EUST3/4 (69 tumors; 83%) or N1 (46 tumors; 55%). FDG-PET after preoperative CRT identified pathologic responders but failed to rule out microscopic residual tumor in 13 of 73 cases (18%). Pathologic response was found to correlate with the post-CRT FDG-PET standardized uptake value (SUV) (P = 0.03) and a post-CRT FDG-PET SUV of ≥ 4 was found to be the only preoperative factor to correlate with decreased survival (2-year survival rate of 33% vs. 60%; P = 0.01). On univariate Cox regression analysis, only post-CRT FDG-PET was found to be correlated with post-CRT survival (P = 0.04).
CONCLUSIONS
Post-CRT FDG-PET was found to be predictive of pathologic response and survival in patients with esophageal carcinoma who undergo preoperative CRT. Esophagectomy should still be considered even if the post-CRT FDG-PET scan is normal because microscopic residual disease cannot be ruled out. Cancer 2004. © 2004 American Cancer Society.
Multimodality therapy with chemotherapeutic agents, radiation therapy, and surgery has increasingly been used to treat locoregionally advanced esophageal carcinoma because of the poor longterm survival of patients treated with surgery alone.1-6 Multimodality therapy, however, is often associated with significant cost and morbidity and appears to have no clear survival benefit, except in a subset of patients who respond to induction chemoradiation therapy (CRT).4, 7-10 Being able to identify this subset of patients has implications not only for the clinical management of esophageal carcinoma patients undergoing multimodality therapy but also for their prognostic stratification. Several preliminary studies have suggested that 2-fluoro-2-deoxy-d-glucose positron emission tomography (FDG-PET) can identify this subset of patients.11-13 To test this hypothesis, we reviewed our recent experience with FDG-PET after CRT but before surgical tumor resection in patients with esophageal carcinoma for the purpose of evaluating the ability of FDG-PET to predict pathologic response and overall survival.
MATERIALS AND METHODS
Patients
We retrospectively evaluated 113 consecutive patients with primary esophageal carcinoma who underwent FDG-PET at the study institution between May 2001 and May 2003. The aim was to evaluate the ability of FDG-PET to predict response and long-term survival after preoperative CRT and tumor resection. Therefore, only those patients who underwent preoperative CRT, FDG-PET, and subsequent surgical tumor resection were included (n = 83). Excluded were 18 patients who did not undergo preoperative CRT and 12 patients who did not undergo surgical tumor resection.
Preoperative Staging and Treatment
All 83 patients included in the current study were required to have histologically diagnosed adenocarcinoma or squamous cell carcinoma of the esophagus before beginning multimodality therapy. Endoscopic ultrasonography (EUS), computed tomography (CT) of the chest and abdomen, barium swallow radiography, and PET were performed to determine pretreatment clinical stage. Patients whose tumors were considered to be resectable but at a high risk of recurrence because of local extension or possible lymph node metastases were eligible to receive preoperative CRT. All patients were assessed by a multidisciplinary team (thoracic surgeon, medical oncologist, radiation oncologist, gastroenterologist, and radiologist) to determine the resectability of their tumor and their physiologic ability to tolerate preoperative CRT and surgery. Acceptable criteria for preoperative CRT and surgery included clinical Stage IIB through IVA disease (according to the American Joint Committee on Cancer [AJCC] TNM staging system), age < 80 years, physiologic ability to tolerate esophagectomy as determined by the individual surgeon, and patient consent to treatment. This study was approved by the institutional review board of the M. D. Anderson Cancer Center.
Treatment Plan
All patients who were candidates for preoperative CRT were treated either according to institutional protocols (n = 47) with induction chemotherapy followed by concurrent CRT or off protocol (n = 36) with concurrent CRT alone. Patients treated on institutional protocols received up to 2 cycles of irinotecan (45 mg/m2), docetaxel (33 mg/m2), and 5-fluorouracil (2 g/m2 as a 24-hour infusion). This was followed by concurrent radiation therapy (50.4 grays [Gy] in 28 fractions) and irinotecan (30 mg/m2/week for 5 weeks), docetaxel (20 mg/m2/week for 5 weeks), and 5-fluorouracil (300 mg/m2/day). Patients treated off protocol received concurrent CRT, comprised of radiation therapy (50.4 Gy in 28 fractions), and chemotherapy, comprised of either cisplatin and 5-fluorouracil or paclitaxel and carboplatin.
At the time of completion of the CRT, tumors were clinically restaged using CT of the chest and abdomen, esophagoscopy with or without EUS, and FDG-PET. If a patient had not developed systemic metastases and could physiologically tolerate surgical resection, an esophagectomy was performed 4–8 weeks after the completion of CRT. The type of esophagectomy performed was dependent on the location of the tumor and the preference of the individual surgeon. Either a transthoracic approach (Ivor-Lewis [two-field] or total [three-field]) or a transhiatal approach to resection was used. The mediastinal and celiac lymph nodes were resected in all patients who underwent surgery.
Positron Emission Tomography
FDG-PET images were obtained on a Siemens HR+ tomograph (Siemens Medical Systems, Hoffman Estates, IL) (resolution, 4.5 mm × 4.5 mm × 4.5 mm; full width, half maximum [FWHM]). All patients fasted for at least 4 hours before undergoing PET. Imaging was performed 45–60 minutes after intravenous injection of FDG (15-20 millicuries [mCi]). The area imaged extended from the mid-neck to the upper thighs. Transmission scanning was performed over the chest and abdomen from five or six overlapping bed positions and, using a rotating germanium-68 pin source, corresponding emission images also were obtained. Transmission scans were acquired at 3 minutes per bed position and emission scans were acquired at 5 minutes per bed position. The PET images were reconstructed using software running standard vendor-provided reconstruction algorithms, and by ordered subset expectation maximization (OSEM reconstruction: two iterations, eight subsets followed by smoothing with a Gaussian filter). Emission data were corrected for scatter, random events, and dead-time losses using Siemens's software, and images were reconstructed both with and without attenuation correction. Images were analyzed visually and quantitatively to determine the standardized uptake value (SUV) for local malignancy and metastases. Images were not corrected for lean body mass and maximal pixel value was utilized. To reduce errors in the SUV measurements, standard calibrations as recommended by the vendor were performed: 68-Ge phantom cylinder calibration, daily uniformity/reference scans, two-dimension and three-dimension normalization, and singles attenuation calibrations as well as monthly detector calibrations. To determine the sensitivity, specificity, and accuracy of FDG-PET in identifying residual disease after preoperative CRT and before esophagectomy, PET images were interpreted retrospectively by an experienced nuclear physician (H.M.) and radiologist (J.E.) who were blinded to all clinical data and whose findings were recorded by consensus. Images were reviewed in all standard planes along with maximum intensity projection images. PET scans were reviewed in combination with CT scans whenever possible. On visual analysis, an abnormality was defined as any instance in which FDG uptake was substantially greater than mediastinal blood pool activity on the attenuation-corrected images. The intensity of FDG uptake in the primary tumor, locoregional lymph node involvement, and distant metastatic disease were assessed as present or absent. A pixilated region of interest (ROI) was outlined within any region of increased FDG uptake and, after correction for radioactive decay, analyzed semiquantitatively according to the following formula: SUV ratio = mean ROI activity (mCi/mL)/injected dose (mCi)/body weight (g)
Treatment Response Criteria
Resected specimens underwent routine histopathologic examination and also were reviewed by a single pathologist (T-T.W.) for percent residual tumor and classification according to three types of pathologic response known to predict long-term survival10: complete response (no residual tumor), microscopic residual disease (< 10% residual tumor), or macroscopic residual disease (> 10% residual tumor). Strict anatomic correlation between percent tumor residual and the SUV value was not performed. All patients were assigned a TNM classification and staged according to the AJCC's 2002 guidelines for pathologic and clinical staging (as determined using EUS, CT, and PET before preoperative CRT).14
Short-term outcome after surgery was assessed by a collection of prospective data regarding length of ventilator dependence, days spent in the intensive care unit, days spent in the hospital, anastomotic leak rates, complications, and death before discharge (within 30 days after surgery). Long-term outcome (i.e., overall survival) was assessed by collecting prospective data or conducting phone interviews until death or the end of the current study period.
Statistical Analysis
Survival probability analyses were performed using the Kaplan–Meier method. Survival was calculated from the date of surgery to the date of death or most recent follow-up contact. Analysis of survival excluded patients who died of noncancer-related causes within 30 days after surgery. Statistical significance was assessed by the log-rank test. Independent predictive factors for survival were determined using Cox regression analysis. Univariate analyses were performed using chi-square analysis. Two-tailed P values of ≤ 0.05 were considered significant. Data analysis was performed by our departmental biostatistician (A.M.C.) with SPSS (SPSS Inc., Chicago, IL) software.
RESULTS
Demographics and Short-Term Outcomes
In the patient population in the current study, there was a predominance of men, adenocarcinomas, and tumors located distally or at the gastroesophageal junction (GEJ) (Table 1). Because all patients underwent preoperative CRT, none had early-stage disease, and many were found to have clinical Stage III or IVA tumors before treatment. Sixty-nine patients (83%) had T3 or T4 disease, 46 (55%) had N1 lymph node metastases, and 6 (7%) had enlarged celiac lymph nodes. Forty-eight patients underwent FDG-PET imaging both before and after preoperative CRT 10 patients underwent FDG-PET imaging only before preoperative CRT, and 25 patients underwent FDG-PET imaging only after preoperative CRT.
| Characteristics | |
|---|---|
| Median age (yrs) (range) | 62 (34–79) |
| Gender | |
| Male | 74 (89%) |
| Female | 9 (11%) |
| Histology | |
| Adeno | 73 (88%) |
| SCCA | 10 (12%) |
| Esophageal location | |
| Upper | 1 (1%) |
| Middle | 5 (6%) |
| Lower | 33 (40%) |
| GEJ | 44 (53%) |
| AJCC TNM clinical stage | |
| IIA | 36 (43%) |
| IIB | 5 (6%) |
| III | 36 (43%) |
| IVA | 6 (7%) |
| PET | |
| Pre-CRT | 10 (12%) |
| Pre - Post-CRT | 48 (58%) |
| Post-CRT | 25 (30%) |
- Adeno: adenocarcinoma; SCCA: squamous cell carcinoma; GEJ: gastroesophageal junction; AJCC: American Joint Commission on Cancer; Pre-CRT: before preoperative chemoradiation therapy; Post-CRT; after preoperative chemoradiation.
The surgical approach was transthoracic in the majority of patients (60%). The median hospital stay was 12 days, and the 30-day surgical mortality rate was 1.2%. Because all patients received preoperative CRT, many demonstrated tumor downstaging; 26 patients had no primary tumor detected after CRT (Table 2). Before CRT, 69 patients (83%) had clinical T3 or T4 disease; at the time of surgery, only 39 patients (47%) were found to have pathologic T3/4 disease after CRT. Before CRT, 46 patients (55%) had clinical N1 disease, whereas after treatment, only 27 patients (32%) were found to have pathologic N1 disease. Of the 71 patients assessed for tumor viability, 54 (68%) demonstrated < 50% tumor viability after preoperative CRT.
| Characteristics | |
|---|---|
| Pathologic T classification | |
| 0 | 23 (28%) |
| 1 | 11 (13%) |
| 2 | 10 (12%) |
| 3 | 35 (42%) |
| 4 | 4 (5%) |
| Pathologic N classification | |
| 0 | 56 (68%) |
| 1 | 27 (32%) |
| Pathologic M classification | |
| 0 | 79 (95%) |
| 1a | 3 (4%) |
| 1b | 1 (1%) |
| AJCC pathologic stage of disease | |
| 0 | 23 (28%) |
| I | 7 (8%) |
| IIA | 24 (29%) |
| IIB | 7 (8%) |
| III | 18 (22%) |
| IVA | 3 (4%) |
| IVB | 1 (1%) |
| Pathologic responsea | |
| Complete response | 26 (31%) |
| Microscopic residual | 27 (33%) |
| Macroscopic residual | 30 (36%) |
| Tumor viability (n = 79)b | |
| > 50–100% | 25 (32%) |
| > 10–50% | 11 (14%) |
| > 0–10% | 23 (29%) |
| 0% | 20 (25%) |
- AJCC: American Joint Committee on Cancer.
- a Pathologic response at the primary tumor: complete response: no viable tumor; microscopic residual: < 10% viable tumor; macroscopic residual: > 10% viable tumor.
- b Tumor viability at the primary tumor was determined in 79 patients by single pathologist who was blinded to outcome (T-T.W.).
Prognostic Utility of FDG-PET
All post-CRT FDG-PET images interpreted as abnormal at the primary tumor had a sensitivity of 82%, a specificity of 29%, and an accuracy of 69% in predicting residual malignancy (Table 3). The false-positive rate for identifying residual malignancy in the region of the primary esophageal tumor was 71%. The false-negative rate for identifying residual malignancy was 18%. As the SUV cutoff value for abnormality was increased (Table 3), the specificity of FDG-PET improved. However, the sensitivity decreased and more patients were found to have viable residual disease despite having normal readings.
| PET criteria | Sensitivityb | Specificityc | Accuracyd |
|---|---|---|---|
| Abnormal primary tumor (No SUV criteria) | 85% | 29% | 69% |
| SUV criteria | |||
| SUV > 2 vs. ≤ 2 | 76% | 19% | 59% |
| SUV > 3 vs. ≤ 3 | 46% | 62% | 51% |
| SUV > 4 vs. ≤ 4 | 26% | 95% | 47% |
| SUV > 5 vs. ≤ 5 | 18% | 95% | 41% |
| SUV > 6 vs. ≤ 6 | 14% | 95% | 38% |
- Post-CRT FDG-PET: postchemoradiation therapy 2-fluoro-2-deoxy-d-glucose positron emission tomography; SUV: standardized uptake value.
- a Residual disease indicates any viable disease in the primary tumor (macroscopic and microscopic residual disease).
- b Sensitivity = true-positive/(true-positive + false-negative).
- c Specificity = true-negative/(true-negative + false-positive).
- d Accuracy = (true-positive + true-negative)/(true-positive + false-positive + true- negative + false-negative).
The pre-CRT FDG-PET SUVs did not appear to correlate with pathologic response after CRT. However, post-CRT FDG-PET SUVs at the primary tumor site did. All tumors in the study measured ≥ 3 cm in size and a correlation was not observed between SUV and tumor size. It is interesting to note that the post-CRT FDG-PET SUV did not distinguish between patients with microscopic residual tumor and patients who had a complete response to therapy (Table 4). To determine whether this was due to the increasing inaccuracy of SUVs with a decreasing percentage of residual malignant cells, we assessed the extent of residual tumor in serial tumor specimens and correlated this value with the post-CRT FDG-PET SUV at the primary tumor site (Fig. 1). Our analysis demonstrated that the post-CRT FDG-PET SUVs decreased as the extent of necrosis increased, until a plateau was reached at < 10% residual tumor (Fig. 1). This observation, together with the inherently poor resolution of PET imaging at low volumes, may explain the inability of post-CRT FDG-PET scans to distinguish microscopic residual disease from a complete pathologic response.
| Complete response (no viable tumor) | Microscopic residual (< 10% viable tumor) | Macroscopic residual (>10% viable tumor) | P valuea | |
|---|---|---|---|---|
| Pre-CRT FDG-PETb (n = 56) | 21 | 17 | 18 | |
| SUVd | 9.6 ± 5.1 | 8.1 ± 3.8 | 10.4 ± 7.2 | 0.46 |
| Post-CRT FDG-PETc (n = 71) | 21 | 25 | 25 | |
| SUVd | 2.8 ± 1.1 | 2.6 ± 1.1 | 6.4 ± 8.1 | 0.01 |
- FDG-PET: 2-fluoro-2-deoxy-d-glucose positron emission tomography; SUV: standardized uptake value; CRT: chemoradiation therapy.
- a P < 0.05.
- b 2-fluoro-2-deoxy-d-glucose positron emission tomography imaging study was performed before preoperative chemoradiation therapy.
- c 2-fluoro-2-deoxy-d-glucose positron emission tomography imaging study was performed after preoperative chemoradiation therapy and before esophagectomy.
- d Standardized uptake value of the primary tumor (mean ± the standard deviation).
Postchemoradiation therapy 2-fluoro-2-deoxy-d-glucose positron emission tomography (PET) standardized uptake value (SUV) (mean, ± the standard error of the mean) of the primary tumor as correlated with the percentage of viable cells in pathologic specimens (n = 68) at the time of esophagectomy. Patients with > 50% tumor viability were found to have a significantly higher average SUV compared with the other patient groups (P = 0.03).
Predictors of Long-Term Outcome
Because post-CRT FDG-PET SUVs were found to correlate with the pathologic response of the primary tumor, we evaluated the ability of post-CRT FDG-PET SUVs to predict long-term survival (Fig. 2). Patients who died 30 days perioperatively were excluded from this analysis. A post-CRT FDG-PET SUV of 4 was utilized as the cutoff value because this was the level at which specificity rose to 95% with few false-positive results (Fig. 1) and at which level pathologic responders were clearly separated from nonresponders (Figs. 3-5, Table 4). The 2-year survival rate was 60% for patients with a post-CRT FDG-PET SUV of < 4 and 34% for those with a post-CRT PET SUV of ≥ 4 (P = 0.01). Univariate Cox regression analysis of various other preoperative clinical predictors of long-term survival (Table 5) demonstrated that only post-CRT FDG-PET SUV was correlated with survival.
Kaplan–Meier analysis of the overall survival of esophageal carcinoma patients according to their postchemoradiation 2-fluoro-2-deoxy-d-glucose positron emission tomography (PET) standardized uptake value (SUV) at the primary tumor site.
| Risk factor | HR | 95% CI | P valuea |
|---|---|---|---|
| Patient age | 1.010 | 0.954–1.068 | 0.734 |
| Gender | |||
| Female | 1.287 | 0.160–10.379 | 0.813 |
| Male | 1 | ||
| Histology | |||
| Squamous cell | 2.876 | 0.593–13.933 | 0.190 |
| Adenocarcinoma | 1 | ||
| Tumor location | |||
| Upper/middle | 1.504 | 0.187–12.099 | 0.701 |
| Lower/GEJ | 1 | ||
| Tumor size | 1.102 | 0.824–1.474 | 0.512 |
| Clinical stage | |||
| III/IVB | 2.371 | 0.610–9.215 | 0.213 |
| IIA/IIB | 1 | ||
| Chemo/RT sequenceb | |||
| C/RT | 3.622 | 0.436–30.060 | 0.233 |
| C→C/RT | 1 | ||
| Pre-CRT FDG-PETc | |||
| SUV > 9.5 | 1.474 | 0.207–10.485 | 0.698 |
| SUV ≤ 9.5 | 1 | ||
| Post-CRT FDG-PETd | |||
| SUV > 4 | 5.036 | 1.251–20.278 | 0.023 |
| SUV ≤ 4 | 1 |
- HR: hazards ratio; 95% CI: 95% confidence interval; GEJ, gastroesophageal junction; chemo/C: chemotherapy; RT: radiation therapy; CRT: chemoradiation therapy.
- a P < 0.05.
- b Induction chemotherapy (C) followed by concurrent radiation therapy (RT) (C→C/RT, “on” protocol); concurrent chemoradiation alone (C/RT, “off” protocol).
- c 2-fluoro-2-deoxy-d-glucose positron emission tomography (FDG-PET) imaging study was performed before preoperative chemoradiation therapy. The standardized uptake value of the primary tumor was determined by radiologists (J.E. and H.M.) who were blinded with regard to patient outcome.
- d 2-fluoro-2-deoxy-d-glucose positron emission tomography (FDG-PET) imaging study was performed after preoperative chemoradiation therapy and prior to esophagectomy. The standardized uptake value of the primary tumor was determined by radiologists (J.E. and H.M.) who were blinded with regard to patient outcome.
No residual carcinoma (SUV 1.6), characterized by ulceration, chronic inflammation, radiation-induced tissue injury, and fibrosis, was noted on (A) the axial noncontrast-enhanced computed tomography (CT) scan, (B) the 2-fluoro-2-deoxy-d-glucose positron emission tomography (FDG-PET) scan, and (C) fusion CT-PET images after chemoradiation therapy. Minimal FDG uptake was noted in the esophagus (cross-hair), correlating with the lack of residual tumor. There was no histologically identifiable residual carcinoma noted on the (D) pathologic specimen (standardized uptake value of 1.6).
Microscopic residual carcinoma (1–10% residual carcinoma, SUV 3.2) was noted on (A) the axial noncontrast-enhanced computed tomography (CT) scan, (B) the 2-fluoro-2-deoxy-d-glucose positron emission tomography (FDG-PET) scan, and (C) fusion CT-PET images after chemoradiation therapy, 1–10% residual carcinoma was noted on the (D) pathologic specimen, characterized by the small microscopic focus of residual adenocarcinoma cells present in the muscularis propria (standardized uptake value of 3.2).
Macroscopic residual carcinoma (> 50% residual carcinoma, SUV 7.5) was noted on (A) the axial noncontrast-enhanced computed tomography (CT) scan, (B) the 2-fluoro-2-deoxy-d-glucose positron emission tomography (FDG-PET) scan, and (C) fusion CT-PET images after chemoradiation therapy, the (D) pathologic specimen demonstrated a > 50% residual carcinoma, as characterized by a macroscopically identifiable tumor mass with a large cluster of residual carcinoma. The example shown here is a squamous cell carcinoma (standardized uptake value of 7.5)
DISCUSSION
The efficacy of preoperative CRT as induction therapy for patients with locoregionally advanced esophageal carcinoma (Stage II, III, or IVA disease) is controversial. To our knowledge, only one of three randomized studies published to date15-17 has shown a statistical survival advantage for its use in patients with esophageal adenocarcinoma. However, all three studies demonstrated improved survival in patients who had achieved a complete or nearly complete pathologic response to induction therapy.4, 7, 9, 10, 15, 16, 18, 19 The ability to identify these responding patients might allow for the more appropriate selection of multimodality treatments for patients who are otherwise treated as a homogenous group. Several preliminary studies in small patient populations dominated by patients with squamous cell carcinomas (24 patients in the study by Brucher et al.12 and 36 patients in the study by Flamen et al.13) or adenocarcinomas (17 patients in the study by Downey et al.11) have suggested that FDG-PET imaging may be useful in this regard. The current study of a large number (n = 83) of predominantly adenocarcinoma patients treated with preoperative CRT and esophageal resection confirms and extends these observations.
The current prospective study has resulted in several important observations. First, even though FDG-PET imaging can identify patients who have a large amount of residual disease (macroscopic residual disease, > 50% viable tumor) after preoperative CRT, it cannot detect a small amount (< 50% viable) of residual tumor (Table 3). It is not clear whether FDG uptake is an indication of tumor cell proliferation or is directly related to the number of viable tumor cells,20, 21 but it does appear that SUV decreases as the percentage of viable cells decreases (Fig. 1). Once the tumor viability is < 50%, the number of tumor cells able to incorporate FDG may be too scarce to allow differentiation of microscopic residual disease from a complete pathologic response. Therefore, even with the most sensitive criteria (Table 3), FDG-PET imaging will falsely identify residual tumor in 18% of the CRT-treated patients (13 of 71 patients). If so, then FDG-PET should not be used as the sole criterion with which to determine whether esophagectomy is needed after preoperative CRT. This recommendation is supported by the study of Nakamura et al.,22 who reported locoregional recurrence in 42% of the patients who were treated with definitive CRT and achieved a complete response according to FDG-PET. Brucher et al. had similar difficulty in using changes in FDG-PET SUV to differentiate between a complete pathologic response and partial, minimal, or subtotal responses in patients with squamous cell carcinoma.12
A second important observation is that FDG-PET imaging could predict both the response to preoperative CRT and a poor prognosis. This ability has been reported in studies of other malignancies, including head and neck carcinoma, lung carcinoma, pancreatic carcinoma, and multiple myeloma.23-26 In the current study, patients who had a post-CRT FDG-PET SUV value of ≥ 4 were found to have a significantly worse 2-year survival rate compared with those with an SUV of < 4 (P = 0.01). In fact, post-CRT FDG-PET was found to be the only preoperative factor we evaluated that could predict survival (Table 5). Other smaller studies in patients with esophageal carcinoma have produced similar preliminary findings.11-13 The results of the current study are statistically significant, even though the follow-up is still relatively short and even though the results may be influenced by the advanced disease observed in and the large number of the patients evaluated. Because to our knowledge no other factor (including histology, tumor size, and clinical stage) can accurately predict prognosis before surgical resection, the importance of these observations cannot be emphasized enough (Table 5). Consequently, the findings of the current study have important therapeutic implications for patients with a poor prognosis who may receive additional therapy prior to surgery. We recently observed that many patients who receive more chemotherapy before concurrent CRT may achieve a better pathologic response and have a better prognosis.10 If FDG-PET imaging could reliably identify CRT response, then multimodality therapy could be tailored to individual patients rather than generalized to a heterogeneous population. This could in turn minimize treatment-related morbidity while maximizing therapeutic benefit.
A third significant observation is that the specificity of FDG-PET imaging after preoperative CRT is poor (29%) unless the SUV cutoff value of the primary tumor is increased to 4, at which point the sensitivity decreases dramatically (18%) (Table 3). The reason for this high false-positive rate may be that the inflammatory changes that follow radiation therapy can lead to falsely elevated SUV values because of the presence of metabolically active leukocytes and macrophages.27-29 The timing of post CRT FDG-PET imaging therefore may be critical because the percentage of false-positive results appears to decrease with time.23 It also is important to realize that SUV values are not strictly quantitative and are dependent on institutional parameters. In the current study, post-CRT FDG PET was performed 4–6 weeks after CRT because of the need to perform a planned esophagectomy during a prescribed time period. It is interesting to note that Weber et al. noted much lower false-positive rates for FDG-PET imaging when it was performed 2 weeks after the initiation of preoperative chemotherapy for esophageal carcinoma (sensitivity of 93% and specificity of 95%).30 Whether the earlier imaging contributed to the lower false-positive rate is unclear because differences in therapeutic management also may have been a factor. Patients in the study by Weber et al. were treated with chemotherapy only (compared with chemotherapy and radiation therapy in the current study), which means that postradiation inflammatory changes may not have occurred.
Several authors have suggested that two FDG-PET scans are needed to predict outcome because this would allow the percent decrease in SUV relative to the baseline to be determined.11-13, 30 However, the results of the current study suggest that a single post-CRT PET study may be sufficient. This issue still should be decided in a multiinstitutional setting in which the SUV cutoff values are determined prior to the trial. It may be that the variance in SUVs and preoperative CRT regimens at different institutions will necessitate two FDG-PET studies for internal standardization. The findings of the current study also suggest that post-CRT PET imaging may be applicable to a broad range of patients treated with a variety of CRT regimens, provided the imaging is performed at a consistent time after CRT (Table 5).
The results of the current study confirm that post-CRT FDG-PET imaging is useful for identifying a subset of esophageal carcinoma patients whose prognosis after preoperative CRT is poor. The SUVs in these patients appear to correlate with the persistence of viable residual disease at the treated site. However, to our knowledge, post-CRT FDG-PET cannot distinguish between patients with microscopic residual disease and complete pathologic responders. Therefore, esophagectomy should remain a therapeutic option even when the post-CRT FDG-PET scan is normal. If future multiinstitution studies confirm these findings, then FDG-PET may some day allow the tailoring of multimodality therapy regimens to individual patients.
Acknowledgements
The authors thank Debbie Smith and Jude Richard for their help in the preparation and review of the article.
