Advancements in Radiation Techniques for Gastric Cancer

Medscape: Continuing Medical Education Online

Adjuvant Chemoradiation in Gastric Cancer

The rationale and benefit of radiotherapy (RT) for managing nonmetastatic gastric and gastroesophageal junction (GEJ) cancers have been well defined by analysis of locoregional failure patterns and clinical trials of its use. Clinically detected locoregional recurrence after surgical resection alone generally exceeds 25%.^1,2 However, as shown by the University of Minnesota reoperation series, local and regional recurrence rates after resection significantly exceed reported rates of clinical recurrence.³ In this study, 107 patients who had undergone surgery with curative intent underwent reoperation to identify early locoregional relapse and attempt surgical resection and salvage. Local disease relapse was seen in 67% of patients, and 88% of these had a component of local or regional failure. In addition, 50% of peritoneal failures were localized to the original tumor bed rather than diffuse throughout the abdomen. High failure rates at the gastric resection bed (54%), anastomosis (26%), or regional nodes (42%) have been corroborated by autopsy series.^4,5

Intergroup 116, a randomized phase III trial of adjuvant postoperative upper abdominal RT and 5-fluorouracil (5-FU)/leucovorin-based chemotherapy for gastric cancer, showed that patients who underwent locoregional cancer therapy had improved overall and disease-free survival (OS/DFS) versus those who underwent surgery alone (3-year OS, 50% vs. 41%; P = .005, and 3-year DFS, 48% vs. 31%; P = .001).⁶ In addition, multiple randomized trials of preoperative chemoradiotherapy (CRT) for esophageal cancer that included tumors of the GEJ^7–9 and a trial exclusively of GEJ tumors¹⁰ have shown improvement in patient survival with locoregional therapy.

Randomized trials of neoadjuvant¹¹ or perioperative chemotherapy¹² without RT have also shown survival improvement compared with resection alone. However, given the modest long-term survival among patients treated with either strategy, the respective gains associated with neoadjuvant chemotherapy or CRT strategies probably present an opportunity to integrate approaches (rather than compete) to improve outcomes, as shown by at least one randomized trial for GEJ cancer.¹⁰

Irradiation Field Design and Delivery

Over the past decade, significant technologic improvements have been made in RT plan design and delivery that are applicable to gastric cancer and show promise in improving patient tolerance and efficacy. Given the large RT target volumes required for treating gastric cancer, the proximity of vital organs in the upper abdomen and lower thorax, and tissue movement from respiration, RT for this disease is both technically and clinically challenging. Intergroup 116, conducted between 1991 and 1998, delivered RT generally using 2-dimensional treatment planning techniques, frequently with opposed anteroposterior/posteroanterior (AP-PA) fields, based on radiographic landmarks and boundaries.^6,13

Although treatment was associated with improved survival, 41% and 32% of patients experienced grade 3 or 4 toxicity, respectively, with 17% unable to complete therapy because of poor tolerance. Minor or major errors in field design were discovered in 35% of patients during preirradiation quality assurance (QA) review. The upfront QA allowed most of the major or minor deviations to be corrected before the start of irradiation, and resulted in only a 6.5% final major deviation rate. Use of upfront QA may have been a key factor in achieving a positive survival advantage for adjuvant CRT in Intergroup 116. Improvements in treatment accuracy and conformality of RT dose present a significant opportunity in the field of radiation oncology to further advance patient outcome.

Improved Accuracy

The results of Intergroup 116 not only showed a benefit in adjuvant CRT for patients with resected gastric cancer but also highlighted that many radiation oncologists are unfamiliar with treating large upper abdominal fields. As a result, the investigators published multiple excellent reviews and articles educating the radiation oncology community on proper field design based on pathways of tumor spread and recurrence patterns in addition to limitations of normal organ exposure.^13–17 Over the past decade, a shift has occurred toward more sophisticated treatment techniques that use multiple and often non-coplanar beams, such as 3-dimensional conformal RT (3DRT) and intensity-modulated RT (IMRT), to treat most cancers, including gastric and GEJ cancer. Increased emphasis has also been placed on accurate 3-dimensional target delineation based on CT anatomy¹⁸ rather than only 2-dimensional field design.

In addition to refinements in RT target definition based on CT-planning, technologic advances have allowed study of and solutions for variability in target and normal organ location during a treatment course (day-to-day interfraction variability) and actual treatment (intrafraction variability caused by respiration). Interfraction variability in stomach location can be substantial, particularly among patients treated with neoadjuvant CRT, because of daily variations in gastric filling.^19,20 Intrafraction changes in target shape (deformation) and location are asymmetric and mainly from respiratory motion. Movement, particularly in the cranial-caudad dimension, frequently exceeds 1 to 1.5 cm.^19–21

To account for this inter- and intrafraction variability, some investigators have suggested treatment margins of up to 3 to 5 cm, depending on the dimension, to accommodate the wide range of variability seen among studied patients.¹⁹ Although using larger treatment margins may ensure more consistent treatment of the intended volumes, they are associated with additional toxicity from excessive coverage in many patients. Image-guided RT (IGRT) allows overexpansion of treatment margins to be minimized through use of daily imaged-guidance (ultrasound, kilovoltage x-ray, or cone-beam CT) to optimize patient setup and therefore accuracy of treatment delivery. Initial experiences with IGRT for gastric cancer were with ultrasound (B-mode acquisition and targeting)^22,23 using upper abdominal vasculature and kidneys and liver for treatment localization. Although not widely adopted, this technique was found to be more accurate than the traditional use of skin marks and to have improved treatment accuracy in most patients.²³

Daily kilovoltage radiograph matching has also been described²⁴ but only improves daily alignment based on boney anatomy rather than organ or nodal location. Perhaps the most significant advance in IGRT technology for the upper abdomen is cone-beam CT. At the time of treatment, a CT is generated by the actual treatment machine, thereby allowing corrections to be made in treatment delivery based on CT anatomy.²³

In addition to daily IGRT to account for interfraction changes in target volumes, intrafraction changes from respiration can be individually accounted for by 4-dimensional CT (4DCT) at the time of simulation. A 4DCT images the patient in all phases of respiration rather than during a single breath-hold. The multiple image sets can then be reconstructed to view and measure tumor and organ motion from respiration (similar to fluoroscopy, but using CT). Target volumes are adjusted during treatment planning to account for their movement range during the complete respiratory cycle. Thus, this technique leads to an increase in the average tumor volume compared with static imaging.²¹ However, given the high variability in target location and movement among individuals, this technology may decrease the irradiated area compared with treatment planning without 4DCT, when large margins are generally applied to account for the range in movement among groups of patients.

Finally, further reductions in the target volume can be accurately devised by using respiratory gating.²⁵ Rather than expanding target volume to account for full respiratory range, this technique accurately limits the irradiated volume to a more stationary phase of the respiratory cycle (usually end-expiration).

The clinical benefit of these technologic advances in treatment accuracy have not been and may not be prospectively compared with conventional 2-dimensional RT, as used in the Intergroup 116 trial. Regardless, the true locoregional control rate is arguably still suboptimal with traditional doses (45 Gy) and techniques. Adjuvant CRT is associated with 15% to 19% clinical local failure rates (often as first site of failure).^6,26,27 Although these rates are an improvement compared with 22% to 29% local recurrence among patients treated with surgery alone in the same series, both rates probably greatly underestimate the locoregional component of disease relapse for which RT impacts on patient outcome. Detection of locally recurrent disease is likely underappreciated in clinical trials and continues to be challenging in the postoperative patient, even with improved imaging. These advances in treatment accuracy are anticipated to continue to contribute to reductions in locoregional failure. In addition, patient tolerance of therapy may improve through reducing unnecessary exposure of normal organs to RT.

Improved Dosimetry

In addition to solutions to improve the accuracy of RT for gastric cancer, advances in the delivery of radiation have enabled significant improvements in the distribution of radiation dose (dosimetry) that conforms to the target volume while reducing exposure to adjacent normal tissues (lung, heart, liver, spinal cord, bowel, and kidneys). Two-dimensional treatment planning, usually with opposed AP-PA fields, can expose a significant volume of kidney to therapy. At least 2 series have described a gradual decline in renal function occurring at least 18 to 24 months after postoperative RT for gastric cancer.^28,29 Although the detected decline in renal function with limited follow-up was not clinically significant, long-term renal function among gastric cancer survivors has not been reported, and improvements in renal sparing may allow delivery of more aggressive chemotherapy regimens or nephrotoxic drugs (e.g., cisplatin). In addition, the total dose using 2-field techniques is limited to a fairly moderate adjuvant dose of only 45 Gy because of spinal cord and small bowel limitations.

Using 3DRT, a patient's abdominal and thoracic anatomy (target and normal structures) are reconstructed in 3 dimensions after CT imaging, allowing optimal field shape, orientation, and relative weighting to be devised virtually. Multiple studies comparing 3DRT techniques with standard 2-field treatment show significant reduction in kidney exposure to radiation (especially the left) and in spinal cord dose, and improved radiation coverage of the intended target volume (see Table 1).^30–33

IMRT represents a further advance in the delivery of highly conformal RT to the upper abdomen. This technique builds on the advantages of 3DRT but incorporates differentially modulated subportions of each beam to further improve dose conformality and gradients. Initially implemented for small target volumes (e.g., prostate cancer), significant dosimetric gains can be made with this modality for the large and complex volumes associated with gastric cancer (see Figure 1). An initial potential disadvantage of IMRT in the upper abdomen was dose inhomogeneity leading to potential “hot spots,” possibly in bowel.³⁴ This limitation is more easily resolved with newer IMRT planning systems.

Multiple comparative dosimetric studies have been conducted comparing IMRT with 2-field plans and 3DRT.^35–37 IMRT is consistently associated with a significant reduction in dose to the highest exposed kidney (usually the left).^{25,34,38–41} In addition, reductions in dose to the liver,^34,37,42 heart,³⁶ lung,⁴³ and spinal cord^39,41 have been reported with IMRT compared with 2-field plans and 3DRT. Furthermore, a comparison study of 3DRT and IMRT treatment plans performed by blinded radiation oncologists showed preference for IMRT 89% of the time, with 90% concordance in choice among physicians.³⁶ Finally, there is further promise of improved radiation dose conformality with the most recent advances in IMRT delivery, such as volumetric modulated arc therapy or helical tomotherapy, in which dose is delivered with a modulated RT beam in a rotational arc. At least one study has shown improved target coverage and dose homogeneity with tomotherapy compared with IMRT with static beam configurations.⁴⁴

With all of these treatment planning systems, especially IMRT, the dosimetric outcomes are highly dependent on not only the technology but also the planner's complex prioritization of the often-competing goals of dose homogeneity, conformality of dose to the target, and limits to normal organ exposure. Thus, the overall reported benefit of these advanced technologies in each study depends on not only technologic advancements but also preferences in how these tools are used to develop an optimal plan. Regardless, the literature shows a stepwise improvement in dose delivery and sparing of abdominal organs with the progressive implementation of these advanced treatment systems.

Though objective evidence of organ sparing has been consistently reported, the clinical significance of these improvements, especially for the kidneys, has been challenged, especially for patients with adequate functional reserve.³⁹ However, the potential benefit of these techniques in sparing normal organs is probably not yet fully appreciated, because Intergroup 116 and most dosimetric studies presented in this article involve patients treated with only 45 Gy of radiation. Although this dose is associated with improvement in patient OS, presumably through reduction in locoregional recurrence, 45 Gy is certainly at the lowest level used in the neoadjuvant or adjuvant treatment of almost all carcinomas and sarcomas.

The benefits of improved organ-sparing with 3DRT, and especially IMRT, may be more apparent with the use of more aggressive chemotherapy regimens. With more advanced radiation techniques, acceptable toxicity has been reported with the concurrent use of oxaliplatin and capecitabine,⁴¹ cisplatin and 5-FU,⁴⁵ or adjuvant cisplatin, taxol, and 5-FU before CRT (with 5-FU).²⁹

Table 1

Dosimetric Comparisons of Advanced Radiotherapy Techniques in the Postoperative Management of Gastric Cancer (45 Gy)

View Table

Irradiation Dose Escalation

Given the likely higher micrometastatic burden remaining in the gastric bed and lymphatics in this disease compared with many other cancers, a reasonable argument can be made that the current standard dose of 45 Gy is likely the minimum effective dose. Therefore, therapeutic gain may be possible through careful dose escalation to selected areas of the target volume, while sparing dose-limiting organs or structures with 3DRT or IMRT techniques. Again, based on the high rates of recurrent locoregional disease in the University of Minnesota experience,³ local control probably has not been convincingly maximized in this disease even with adjuvant CRT. Because conventional techniques have been limited to 45 Gy, dose escalation can be evaluated using IMRT for patients with highest risk for microscopic residual after resection (e.g., positive margins, extensive nodal involvement, limited nodal dissection).

Dosimetric comparison between 2-field (A), 3-dimensional radiotherapy (B), and intensity-modulated radiotherapy plans (C) in a patient receiving neoadjuvant chemoradiotherapy. Colored lines represent 45- (red), 43- (yellow), 40- (green), 35- (blue), and 25-Gy (cyan) isodose lines.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 8, 4; 10.6004/jnccn.2010.0032

View Full Size

Dosimetric comparison between 2-field (A), 3-dimensional radiotherapy (B), and intensity-modulated radiotherapy plans (C) in a patient receiving neoadjuvant chemoradiotherapy. Colored lines represent 45- (red), 43- (yellow), 40- (green), 35- (blue), and 25-Gy (cyan) isodose lines.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 8, 4; 10.6004/jnccn.2010.0032

Dosimetric comparison between 2-field (A), 3-dimensional radiotherapy (B), and intensity-modulated radiotherapy plans (C) in a patient receiving neoadjuvant chemoradiotherapy. Colored lines represent 45- (red), 43- (yellow), 40- (green), 35- (blue), and 25-Gy (cyan) isodose lines.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 8, 4; 10.6004/jnccn.2010.0032

At least 2 series have reported evidence of improved locoregional control with dose escalation in the postoperative setting. A retrospective review of 63 patients treated with postoperative CRT at the Mayo Clinic Rochester noted higher locoregional control with doses of RT greater than 50 Gy.⁴⁶ Another encouraging study was reported by Arcangeli et al.,⁴⁷ describing the outcome of 40 patients treated postoperatively with hyperfractionated RT (1.1 Gy twice daily) to an escalated dose of 55 Gy with concurrent 5-FU. Acute grade 2 or higher hematologic toxicity, nausea and vomiting, or diarrhea was only experienced in 23%, 23%, and 3% of patients, respectively, with no significant late toxicity observed. With a median follow-up of more than 5 years, the in-field recurrence rate was only 7.5%, and 52% of patients were alive. Although neither institution used IMRT in their series, these data support the premise that increase doses of RT beyond 45 Gy may improve patient outcome. The dosimetric advantages of IMRT certainly facilitate the further investigation this strategy.

Clinical Outcomes with Advanced Technologies

Although prospective randomized studies have not been performed to compare advanced radiation techniques versus traditional techniques, and likely will not, multiple single-institution series have observed improved tolerance of gastric RT and favorable clinical outcomes with these newer techniques. In the clinical series from Mayo Clinic Rochester, grade 4/5 complications were observed in 22% of patients treated with a 2-field technique (AP-PA) compared with only 4% among patients whose treatment included 4 or more fields (3DRT; P = .045).⁴⁶ Another series of 82 patients treated with 3DRT (5-field with concurrent 5-FU) from the Princess Margaret Hospital reported similar toxicity and outcomes to Intergroup 116 with 57% of patients experiencing a grade-3 or more toxicity.⁴⁵

In their initial clinical experience with IMRT in the postoperative treatment of gastric cancer, Milano et al.³⁸ treated 6 patients with 50.4 Gy with 5-FU and saw no acute grade 3 toxicity. In a subsequent larger series, Boda-Heggemann et al.⁴¹ compared a 27-patient cohort who received 3DRT with 5-FU and leucovorin with 33 patients treated with IMRT and capecitabine and oxaliplatin. Patients treated with IMRT had a higher completion rate of treatment, less rise in serum creatinine in the year after therapy, and a 20% absolute improvement in 2-year OS.

Conclusions

Although significant progress has been made in improving the outcome of patients with resectable gastric or GEJ cancer using adjuvant chemotherapy and radiation, long-term survival rates and toxicity from therapy are not yet satisfactory. Technologic improvements in the accuracy of RT delivery through CT-based treatment planning with well-defined target volumes, 4-dimensional treatment planning, and IGRT have shown improvement over traditional techniques used only a decade ago. In addition, the improved conformality of dose delivery with 3DRT and IMRT has been shown to reduce the dose to critical organs such as the kidney, liver, spinal cord, and heart, possibly reducing morbidity and facilitating intensification of chemotherapy and chemoradiation regimens. Furthermore, these advances in both accuracy and dose delivery afford the opportunity of radiation dose escalation to maximize locoregional control. As these advances in treatment accuracy and delivery continue to emerge and be refined, their clinical application warrants continued study.