Individualized Radiation Dose Escalation Based on the Decrease in Tumor FDG Uptake and Normal Tissue Constraints Improve Survival in Patients With Esophageal Carcinoma

Abstract

Background:

To determine whether individualized radiation dose escalation after planned chemoradiation based on the decrease in tumor and normal tissue constraints can improve survival in patients with esophageal carcinoma.

Methods:

From August 2005 to December 2010, 112 patients with squamous esophageal carcinoma were treated with radical concurrent chemoradiation. Patients received positron emission tomography-computer tomography scan twice, before radiation and after radiation dose of 50.4 Gy. All patients were noncomplete metabolic response groups according to the Response Evaluation Criteria in solid tumors. Only 52 patients with noncomplete metabolic response received individualized dose escalation based on tumor and normal tissue constraints. Survival and treatment failure were observed and analyzed using SPSS (13.0).

Results:

The rate of complete metabolic response for patients with noncomplete metabolic response after dose escalation reached 17.3% (9 of 52). The 2-year overall survival rates for patients with noncomplete metabolic response in the conventional and dose-escalation groups were 20.5% and 42.8%, respectively(P = .001). The 2-year local control rates for patients were 35.7% and 76.2%, respectively (P = .002). When patients were classified into partial metabolic response and no metabolic response, 2-year overall survival rates for patients with partial metabolic response were significantly different in conventional and dose-escalation groups (33.8% vs 78.4%; P = .000). The 2-year overall survival rates for patients with no metabolic response in two groups (8.6% vs 15.1%) did not significantly differ (P = .917).

Conclusion:

Individualized radiation dose escalation has the potential to improve survival in patients with esophageal carcinoma according to increased rate of complete metabolic response. However, further trials are needed to confirm this and to identify patients who may benefit from dose escalation.

Introduction

Concurrent chemoradiation is the standard therapy for locally advanced esophageal cancer with recommended thoracic radiation dose of 50.4 Gy¹, but the 5-year actuarial survival is approximately 25%.^1,2 Radiation dose escalation is an optional method for improving patients’ survival. In the INT0123 trial, dose escalation from 50.4 Gy to 64.8 Gy did not increase overall survival rate,² which could be due to higher toxicities and no therapeutic gain in dose escalation for pathological complete response (pCR) after planned radiation. Our previous study indicated that only some patients benefited from a late course accelerated hyperfractionated radiation schedule.³ We speculate that if patients with pCR did not receive dose escalation after the standard radiation of 50.4 Gy, no patient would have died of treatment-related toxicities.

An earlier study has shown that there is a positive correlation between patients’ survival time and the scale of postradiation histopathologic response in patients with esophageal cancer after chemoradiation.⁴ Patients with pCR had a 5-year overall survival rate of 61.6%, which is higher than the survival rate for patients with pathological incomplete or nonresponse.⁴ We postulated that if more patients with pathological incomplete or nonresponse after planned radiation could achieve pCR via radiation dose escalation, prognosis would be better. This scheme would avoid an increase in toxicity and a decrease in survival caused by dose escalation in patients with pCR.

Metabolic response (MR) using positron emission tomography-computer tomography (PET-CT) is a surrogate for histopathologic response in predicting sensitivity to treatments of patients with esophageal cancer.^5–7 In previous clinical trials, tumor response has been evaluated by measuring the decrease in FDG uptake according to the European organization for Research and Treatment of Cancer (EORTC) criteria.⁸ To achieve more pCR in esophageal carcinoma radiotherapy without increasing toxicity, individualized dose escalation could be selected based on the decrease in tumor FDG uptake and normal tissue constraints. First of all, it is necessary to determine the feasibility of individualized radiation dose escalation after planned chemoradiation based on the decrease in tumor FDG uptake and normal tissue constraints, which is the aim of the present study.

Materials and Methods

Patient Population

From August 2005 to December 2010, all patients with squamous esophageal carcinoma were treated with radical concurrent chemoradiation and received PET-CT scan twice before radiation and after radiation dose of 50.4 Gy at Yantai Yuhuangding Hospital (Yantai, China). Recruitment criteria were as follows: patients with noncomplete metabolic response (non-CMR) after radiation dose of 50.4 Gy and survival time of more than 5 months were enrolled in the study according to the EORTC criteria,^8,9 which differs from previously published clinical trials. Patients without narrow lumen after planned radiation received esophagoscope examination to exclude those with complete response after planned radiation. All tumors were restaged according to the TNM staging system of the 2007 International Union against cancer on cancer staging system¹⁰ based on physical examination and radiographic images, without ultrasound endoscope. Patients would be excluded if distant metastasis other than supraclavicular lymph nodes and serious medical diseases were found before concurrent chemoradiation and grade IV toxicity occurred after radiation dose of 50.4 Gy. Two groups of patients were recruited according to clinical requirement and patients’ will. Fifty-two patients received radiation dose escalation, whereas the other patients were treated with conventional radiation dose of 50.4 Gy. Patients’ characteristics for different fractionated dose are summarized in Tables 1 and 2. Karnofsky performance status was ≥70 for all patients. Pretreatment examinations included medical history and physical examination, complete blood cell count, electrocardiogram, chest radiograph, esophageal barium-swallow imaging, esophagoscopy, chest PET-CT scan, bone scan, and ultrasonic examination for abdomen, including liver, kidney, spleen, and retroperitoneal lymph nodes.

Table 1.

Patient Characteristics.

	Conventional Radiation	Dose Escalation	n (%)
Age, years
<65	48	41	89 (79.5)
≥65	12	11	23 (20.5)
Median	62	59	62
Range	45-70	41-67
Gender
Male	43	37	80 (71.4)
Female	17	15	32 (28.6)
T stage
T2	1	1	2 (1.8)
T3	33	30	63 (56.2)
T4	26	21	47 (42.0)
N stage
N0	41	40	81 (72.3)
N1	19	12	31 (27.7)
Site
Upper thoracic	14	7	21 (18.8)
Middle thoracic	38	41	79 (70.5)
Lower thoracic	8	4	12 (10.7)
Length, cm
≤5	17	12	29 (25.9)
>5	43	40	83 (74.1)

Abbreviations: N, node; T, tumor.

Table 2.

Treatment Characteristics and Total MR.

	Conventional Radiation (N = 60)	Dose Escalation (N = 52)
Radiation dose, Gy
Range (median)	50.4 (50.4)	63.0-70.0 (66.6)
PMR	46.7% (28/60)	50.0% (26/52)
NMR	53.3% (32/60)	50.0% (26/52)

Abbreviations: MR, metabolic response; NMR, nonmetabolic response; PMR: partial metabolic response.

The PET/CT Examination

First, ¹⁸F-FDG PET/CT scan was performed before concurrent chemoradiation (preradiation). A radiation plan was designed after initial ¹⁸F-FDG PET/CT scan. For patients who received radiation dose escalation, a second ¹⁸F-FDG PET/CT scan was used immediately after planned radiation of 50.4 Gy (during radiation). The third FDG-PET/CT scan was acquired 1 week after completion of fractionated radiation (postradiation).

Discovery LS PET-CT (General Electric Co, Milwaukee, USA Wisconsin) was used to scan patients in the Department of Nuclear Medicine at Yantai Yuhuangding Hospital (Yantai, China). The FDG came from PET/CT Center at Yantai Yuhuangding Hospital at a mean dosage of 350 (range: 259-444) MBq. The PET scanning range was from the angulus mandibulae level to the lower border of the second lumbar vertebra with a slice thickness of 5 mm.

The PET images were reconstructed using ordered subsets expectation maximization algorithm in a 128 × 128 matrix. A semiquantitative analysis using attenuation-corrected images was performed in all slices imaging the primary tumor. This was based on the calculation of the SUVmax, which is defined as the maximum tracer uptake in the lesion relative to the injected dose and body weight, according to the following formula: SUV = tissue activity concentration (Bq/kg)/injected dose (Bq)/body weight (kg). Primary tumor and all areas of the invasions were considered as regions of interest.

According to the EORTC criteria, complete metabolic response (CMR) was defined as complete resolution of ¹⁸F-FDG uptake within the tumor volume so that it was indistinguishable from surrounding normal tissue.^8,9 Patients with CMR were excluded from the study.

Treatments

Conventional radiation

The computed tomography (CT) images were obtained from the angulus mandibulae level to the lower border of the second lumbar vertebra on PET-CT scan. The images were then transferred to the 3-dimensional planning system (ADAC Pinnacle 1.5, Philips, Philadelphia, USA Pennsylvania). The delineation of clinical target volumes was based on CT, barium esophagogram, endoscopic examination, and PET imaging. The clinical target volumes were a 3-cm proximal and distal margin, and a 0.5 to 0.8 cm radial margin was added to the Gross Tumor Volume (GTV). The planning target volume encompassed a 1-cm proximal and distal margin and a 0.5-cm radial margin based on clinical target volumes. Sixty patients received conventional radiation scheme. The scheme of conventional radiation was a total dose of 50.4 Gy at 180 cGy per fraction once daily in 5.5 weeks according to previous reports.^1,2

Dose escalation

Fifty-two patients with non-CMR received dose escalation after conventional radiation dose of 50.4 Gy with conventional fractions schedules and conformal technique (Table 2). Escalated dose was from 10 Gy to 20 Gy, the total dose of which was set to less than 70 Gy for esophagus tolerance.¹¹

Dose limitation of normal organ

The maximum spinal cord dose was 4500 cGy or lower. The volumetric percentage of the whole lungs that received a radiation dose of at least 2000 cGy was less than 30% in order to decrease the risk of severe complications. The average dose for each lung is 1800 cGy or lower. Treatments were designed using computerized radiation dosimetry and delivered by 6-MV X-rays from a linear accelerator (Varian Clinac 2300 C/D; Varian).

Chemotherapy

Chemotherapy scheme was 5-FU 750 mg/m² (Qilu Pharmaceutical Co, Ltd, China) d1-5 and cisplatin75mg/m² (Qilu Pharmaceutical Co, Ltd) d1-3 by intravenous infusion every 3 weeks. In the case of neutropenia, recombinant human granulocyte colony-stimulating factors (Qilu Pharmaceutical Co, Ltd) were used to ensure the completion of chemoradiation.

Toxicities and treatment-related death

Patients were examined weekly throughout the chemoradiation or more often if clinically indicated by medical history, physical examination results, or toxicity values. Liver and renal function tests were performed before and after each chemotherapy cycle. Acute and late radiation toxicities were graded according to the Radiation Therapy Oncology Group (RTOG) criteria and Common Toxicity Criteria version 4.0.^12,13 Treatment-related deaths were recorded.

Follow-Up Evaluation

After the completion of treatments, patients were evaluated by physical examination, esophageal barium tomography, and CT scan at 3-month intervals for 2 years and at 6-month intervals thereafter. Biopsy was required at the esophageal carcinoma site at 3 months and at the time of any X-ray evidence of local recurrence. Survival times were calculated from the start of radiotherapy to the date of the diagnosis of death or the last follow-up.

Patterns of Failure

Treatment failure analysis was based on local esophagus recurrence, regional lymph node spread or recurrence, and distant metastasis. Local failure was defined as any recurrence of the primary tumor, including persistent disease after initial treatment. Regional lymph node failure was defined as regrowing or newly developed mediastinal or supraclavicular lymphadenopathy, and distant failure included any site beyond the primary tumor and regional lymph nodes.

Study End Points and Data Statistical Analysis

The first primary end point of this study was local control rate and overall survival time for 3 years after patients were stratified. The second primary end point was to observe patterns of treatment failures.

Statistical analyses were carried out using SPSS (version 13.0). Overall survival time and local control time were calculated with the Kaplan-Meier method, and survival curves were compared statistically using the log-rank test.

Results

All patients were followed up according to the trial scheme, with a median of 18 months and the longest follow-up time of 40 months. Two patients were lost to follow-up.

Overall Survival Rate and Local Control Rate

The 1-, 2-, 3-year overall survival rates were 46.7%, 20.5%, and 0% for patients in conventional group and 60.4%, 42.8%, 21.3% for those in dose-escalation group, respectively (P = .001). The median overall survival times were 12.5 months (95% confidence interval [CI] 8.59, 16.42) and 22.8 months (95% CI 12.52, 33.01). The 1-, 2-, 3-year local control rates for patients in conventional and dose-escalation groups were 70.3%, 35.7%, 0% and 89.5%, 76.2%, 68.4%, respectively (P = .002). The median local control times were 18.9 months (95% CI 11.52, 26.22) and 24.0 months (95% CI 20.46, 27.47), respectively.

Stratification Analysis in Overall Survival

Two-year overall survival rates for patients with partial metabolic response (PMR) after planned radiation were significantly different in conventional and dose-escalation groups (33.8% vs 78.4%; P = .000). The median overall survival times were 18.4 months (95% CI 10.55, 26.12) and 35.7 months (95% CI 26.92, 44.55). However, there was no difference in 2-year overall survival rate for patients with no metabolic response (NMR) between the 2 groups (8.6% vs 15.1%, P = .917). The median overall survival times were 9.1 months (95%CI 4.31-13.83) and 8.6 months (95%CI 7.78-9.36).

Toxicity of Treatment and Death

Acute and late radiation toxicities were mainly esophagitis, hematologic toxicity, and radiation pneumonia. There were a total of 3 (5.8%) treatment-related deaths for patients who underwent dose escalation due to fistula, cardiac, or hematologic toxicities.

Patterns of Failure

Persistence of disease and distant failure were the most important causes of treatment failure (Table 3). Despite dose escalation, 13.4% (7 of 52) of patients with persistence of disease after radiation had local failure, and 2 patients had local recurrence. The rate of distant metastasis tended to be higher in patients who underwent conventional therapy (30%) than in dose-escalation group (28.8%; Table 3), although the difference was not statistically significant (P > .05).

Table 3.

Patterns of Failure.

Patterns of Failure	Conventional Group		Dose-Escalation Group
	No.	%	No.	%
Alive/no failure	4	6.7	7	13.5
Total failure	46	76.6	38	73.0
Persistent local disease	11	18.3	7	13.4
Local/regional failure	4	6.7	2	3.8
Distant failure	18	30.0	15	28.8
Local/regional/distant failure	13	21.7	14	26.9
Treatment-related death	4	9.3	3	5.8
Dead of medical disease	6	1.9	4	7.7

Discussion

As of this writing, there are several dose-escalation trials for esophageal carcinoma in literatures^2,14; however, no individuated dose-escalation trial has been reported. To our best knowledge, this is the first clinical trial that adopts individualized radiation dose escalation in esophageal carcinoma treatment, which differs from previously published clinical trials. We chose 50.4 Gy for planned radiation according to previous data because earlier trials have shown that radiation doses below 50 Gy adversely impact overall survival.^1,2 Maximal dose from individualized dose escalation was set to 70 Gy according to the literature¹⁴ to limit toxicities.

Patients with CMR after planned radiation did not receive dose escalation in the current trial, which we believe is important when choosing individualized dose escalation. Otherwise, patients with CMR could have had higher treatment-related toxicities associated with dose escalation. As described in the RTOG9405 trial report,² 10% of patients (11 patients) died from treatment-related toxicities in the high-dose (64.8 Gy) arm as compared with 2% in the standard-dose (50.4 Gy) arm. Of the 11 deaths in high-dose arm, 4 deaths occurred during or after dose escalation beyond 50.4 Gy.² It seems reasonable to assume that for patients with pCR after radiation dose of 50.4 Gy, dose escalation may increase treatment-related toxicities/deaths instead of survival advantage.

We speculate that increased rate of CMR after dose escalation may have contributed to a better survival rate because CMR is known as a positive prognosticator.^15,16 In all, 17.3% (9 of 52) of patients with non-CMR after planned radiation received individualized dose escalation to achieve CMR in the trial. It was observed that dose escalation for patients with non-CMR after planned radiation improved overall survival and local control in the trial (Figures 1 and 2). Escalated dose for patients with non-CMR varied based on normal tissue constraints. For patients with minor residual tumor after planned radiation, lower dose escalation would achieve CMR without increasing toxicities (Figure 3). These patients could benefit the most from dose escalation.

Figure 1.

Overall survival difference between patients in conventional group and dose-escalation group (P = .001).

Figure 2.

Local control difference between patients in conventional group and dose-escalation group (P = .002).

Figure 3.

Survival difference between patients with PMR and NMR in both groups (P = .000). PMR indicates partial metabolic response; NMR, no metabolic response.

If patients with non-CMR were further stratified as PMR and NMR according to the EORTC criteria in the trial, 3-year survival rates were significantly higher for PMR (78.4%) in the dose-escalation group as compared with that for NMR (33.8%) in the conventional group (Figure 3), which is in agreement with previous reports.^3,16 It has been shown that the higher decrease in tumor FDG uptake after dose escalation, the better the survival of patients with esophageal carcinoma.⁶ This conclusion could be true for patients with PMR but may not be so for patients with NMR. Patients with NMR required a higher radiation dose according to the formula, while higher dose could mean more toxicities that may override the survival advantage considering the S model dose–effect curve.¹⁷ Therefore, the amount of dose escalation is another important factor that could have influenced the survival advantage in the trial.

For patients with esophageal carcinoma having NMR after dose escalation, however, dose escalation could not change the persistence of local disease because of tumor resistance to X-ray. In the RTOG 9405 trial, the persistent local disease did not alleviate in high-dose arm,² which could be partially explained by lacking of effect of dose escalation in patients with NMR after planned radiation. In the trial, dose escalation for patients with NMR after planned radiation did not bring survival benefit (Figure 3). In fact, higher dose escalation could have brought more treatment-related deaths. As such, population was the third factor to consider in dose escalation. This could be the limitations of this clinical trial in terms of treatment-related deaths. The next question is was the conventional fraction scheme adopted in this trial appropriate? In the previous trial, our results indicated that an increase in radiation intensity at the late course did not improve survival advantage for radio-resistant esophageal carcinoma.⁴ Therefore, the optimal treatments for esophageal carcinoma with radiation resistance still require further studies.

In the present trial, PET was merely used to primarily determine the necessity of dose escalation and to calculate individual doses. The accuracy of PET in predicting pathologic response has not yet been investigated. Furthermore, cautions should be taken for higher dose escalation in esophageal carcinoma radiotherapy to avoid increased toxicities.

Limitations

This trial was designed based on MR, which we used to stratify patients during enrollment. However, we understand that MR could not substitute for histopathologic response.

Conclusion

In conclusion, our trial demonstrated that individualized dose escalation has the potential to improve patients’ survival rate as indicated by increased rate of CMR. However, further studies are warranted to confirm this and to identify the characteristics of patients who may benefit from dose escalation.

Abbreviations

CMR

complete metabolic response

MR

metabolic response

NMR

no metabolic response, including PMD and SMD based on EORTC recommendations (8)

non-CMR

noncomplete metabolic response

pCR

pathological complete response

PET-CT

positron emission tomography-computer tomography

PMR

partial metabolic response

RTOG

Radiation Therapy Oncology Group