Radiation Oncology: A Question-Based Review

Gastric Cancer

Steven H. Lin and Salma Jabbour

image Background

What is the incidence of gastric cancer in the U.S. and worldwide?

1.     U.S: 21,130 cases/yr (in 2009), with 10,620 deaths (7th leading)

2.     Worldwide: ~875,000 new cases/yr; 2nd-leading cause of death (behind lung cancer)

Where are the high-incidence areas in the world?

The highest incidences are found in East Asia (Japan and China) > South America > Eastern Europe.

What are some acquired and genetic risk factors for developing gastric cancer?

1.     Acquired factorsHelicobacter pylori infection, high intake of smoked and salted foods, nitrates, diet low in fruits/vegetables, smoking, RT exposure, obesity, Barrett esophagus/GERD, prior subtotal gastrectomy

2.     Genetic factors: E-cadherin (CDH-1 gene) mutation, type A blood group, pernicious anemia, HNPCC, Li-Fraumeni syndrome

How does tumor location relate to the underlying etiology of gastric adenocarcinoma?

Body and antral lesions are associated with H. pylori infection and chronic atrophic gastritis, whereas proximal gastric lesions (gastroesophageal [GE] junction, gastric cardia) are associated with obesity, GERD, and smoking.

Which has poorer prognosis: proximal or distal gastric cancer?

Stage for stage, proximal gastric cancer has a poorer prognosis.

What are the 2 histologic types of gastric adenocarcinoma? How do these 2 types differ in terms of etiology of the gastric cancer?

Intestinal and diffuse are the 2 histologic types of adenocarcinomas. Intestinal type are differentiated cancers with a tendency to form glands, occur in the distal stomach, and arise from precursor lesions seen mostly in endemic areas and in older people, more commonly men, suggesting an environmental etiology. Diffuse type are less differentiated (signet ring cells, mucin producing), have extensive submucosal/distant spread, and tend to be proximal. They do not arise from precancerous lesions, are more common in low-incidence areas, and are more common in women and younger people, suggesting a genetic etiology.

What is the the Borrmann classification of gastric cancer?

The Borrmann classification is based on the gross morphologic appearance. It is divided into 5 types:

1.     Type I: polypoid/fungating

2.     Type II: ulcerating

3.     Type III: ulcerating/infiltrative

4.     Type IV: diffusely infiltrating (linitis plastica)

5.     Type V: cannot be classified (most aggressive)

What are the lymphatic drainages of the stomach? Please also include the Japanese Research Society (JRS) classification of nodal designation.

1.     1st echelon: N1–perigastric nodes (lesser and greater curvature) and periesophageal nodes (proximal gastric)

2.     2nd echelon: N2–celiac axis, common hepatic, splenic

3.     More distant: N3–hepatoduodenal, peripancreatic, mesenteric root; N4–portocaval, PA nodes, middle colic

4.     JRS N1–N4 are not the same as AJCC staging.

What are the patterns of spread for gastric cancer?

Local extension to adjacent organs, lymphatic mets, peritoneal spread, or hematogenous (liver, lung, bone). Liver/lung mets are generally for proximal/GE junction tumors.

What are the anatomic boundaries and organs for the stomach?

1.     Superior: diaphragm, left hepatic lobe

2.     Inferior: transverse colon, mesocolon, greater omentum

3.     Anterior: abdominal wall

4.     Posterior and lateral: spleen, pancreas, left adrenal, left kidney, splenic flexure of colon

What is the most important prognostic factor for gastric cancer?

TNM stage is the most important factor, with the histologic grade and Borrmann types not being independently prognostic apart from tumor stage. However, in general, Borrmann type I and II are more favorable compared to type IV.

Is all nodal involvement equally prognostic for gastric cancer?

No. The # and location of nodes are important. Min LN involvement adjacent to the primary lesion is more favorable.

image Workup/Staging

How do pts with gastric cancer generally present?

Anorexia, abdominal discomfort, weight loss, fatigue, n/v, melena, and weakness from anemia

What aspects of the physical exam are relevant for evaluating a pt for a possible gastric malignancy?

General physical with focus on abdominal mass (local extension), liver mets, ovarian mets (Krukenberg tumor), distant LN mets (Virchow: left SCV; Irish: left axillary; Sister Mary Joseph: umbilical), ascites, Blumer shelf (rectal peritoneal involvement)

What is important in the workup for gastric cancer?

Gastric cancer workup: H&P (onset, duration, Hx of risk factors), CBC, CMP, esophagogastroduodenoscopy + Bx, EUS +/− FNA of regional LN mets, CT C/A/P, and diagnostic laparoscopy to r/o peritoneal seeding

How many layers are seen on EUS when imaging the GI tract?

5 layers are seen on EUS: layers 1, 3, and 5 are hyperechoic (bright), and layers 2 and 4 are hypoechoic (dark). Layer 1 is superficial mucosa, layer 2 is deep mucosa, layer 3 is submucosa, layer 4 is muscularis propria, and layer 5 is subserosa fat and serosa.

What is the rate of upstaging to stage IV using diagnostic laparoscopy?

35%–40% of pts are found to have mets using diagnostic laparoscopy.

Why is PET imaging not routinely used in staging gastric cancer?

In 1 study, only approximately two thirds of primary tumors are FDG avid (Shah et al., Proc ASCO 2007), with GLUT-1 transporter rarely present on the common subtypes of gastric cancer (signet ring and mucinous). Therefore, there are too many false negatives.

What is the AJCC 7th edition (2009) T-staging classification for gastric cancer?

1.     Tis: confined to mucosa without invasion to lamina propria

2.     T1a: invades lamina propria or muscularis mucosae

3.     T1b: invades submucosa

4.     T2: invades muscularis propria

5.     T3*: penetrates subserosa without invasion of visceral peritoneum (serosa)

6.     T4a*: invades serosa

7.     T4b: invades adjacent structures

8.     *Tumor is classified as T3 if it penetrates through the muscularis propria with extension into the gastrocolic or gastrohepatic ligaments, or into the greater or lesser omentum without perforation of the visceral peritoneum covering these structures. Tumor is classified as T4 if it penetrates the visceral peritoneum covering the gastric ligaments or the omentum.

What is the AJCC 7th edition (2009) N-staging classification for gastric cancer?

1.     N1: 1–2 LNs

2.     N2: 3–6 LNs

3.     N3: ≥7 LNs

4.     N3a: 7–15 LNs

5.     N3b: >15 LNs

What are the AJCC 7th edition (2009) stage groupings for gastric cancer?

1.     Stage IA: T1N0 (adds to 1)

2.     Stage IB: T1N1, T2a-bN0 (adds to 2)

3.     Stage IIA: T1N2, T2N1, or T3N0 (adds to 3)

4.     Stage IIB: T2N2, T3N1, T4aN0 (adds to 4)

5.     Stage IIIA: T4aN1, T3N2, T2N3 (adds to 5)

6.     Stage IIIB: T4bN0, T4bN1, T4aN2, T3N3 (adds to 6 mostly)

7.     Stage IIIC: T4bN2, T4aN3, T4bN3 (adds to 7 mostly)

8.     Stage IV: TXNXM1

What is the 5-yr OS for the various stages of gastric cancer?

T1N0, 90%; T2N0, 52%; T3N0, 47%; T4N0, 15%; TXN1, 10%; TXN2, 10%; TXN3, 10%; TXN4, 3% (Japanese data: modified from Noguichi Y et al., Cancer 1989). These results may not reflect outcomes in the U.S. There is a chance that the biology of gastric cancer is different in Japanese cohorts and that gastric cancer screening is used in Japan to discover early stages of Dz. Surgery is probably easier in the Japanese population because of lower BMIs. Better outcomes may also reflect more extensive LND performed as standard practice.

image Treatment/Prognosis

What surgical margin is necessary for resection of gastric cancer?

5 cm proximal and distal margin (except for select cases removed endoscopically)

In what tumor location is subtotal vs. total gastrectomy indicated? Is there a benefit of advocating total gastrectomy for most gastric tumors?

Subtotal gastrectomy for distal tumors (antrum/body); total gastrectomy for proximal tumors (cardia, greater curvature)

No. According to the following 2 trials, there is no benefit of advocating total gastrectomy:

Gouzi et al. randomized distal tumors to total gastrectomy vs. subtotal gastrectomy. There were no differences in morbidity/mortality (1.3% vs. 3.2%) or survival outcomes (5-yr OS 48%). (Ann Surg 1989)

Italian data from a 2nd randomized trial (Bozzetti F et al., Ann Surg 1999) showed no difference in 5-yr survival between subtotal gastrectomy (65%) and total gastrectomy (62%).

Should splenectomy be performed for proximal gastric tumors to get splenic LN clearance?

Avoid splenectomy if possible, since there is no value of splenectomy in a randomized trial (Csendes A et al., Surgery 2002). Splenectomy and pancreatectomy had an adverse impact on survival in the Dutch and MRC D1 vs. D2 RCTs (see below).

How are GE junction cancers classified, and how is the classification important therapeutically?

GE junction cancers are classified by the Siewert classification as 3 entities:

1.     Type I: adenocarcinoma of distal esophagus, arising from Barrett, that infiltrate GE junction

2.     Type II: adenocarcinoma of cardia portion, arising from cardia and short segment of intestinal metaplasia at GE junction

3.     Type III: adenocarcinoma of subcardial stomach, which may infiltrate GE junction or distal esophagus from below

4.     Type I has lymphatic drainage reminiscent of esophageal primaries (mediatinal and celiac), whereas types II–III drain to celiac, splenic, and para-aortic (P-A) nodes. Esophagectomy is typically recommended for type I tumors, whereas gastrectomy is recommended for type II–III tumors.

What are the types of nodal dissections for gastric cancer?

1.     D0: no nodal dissection

2.     D1: perigastric nodes removed

3.     D2: D1 + periarterial nodes (left gastric, hepatic, celiac, splenic)

4.     D3: D2 + hepatoduodenal, peripancreatic, mesenteric root, portocaval, P-A nodes, middle colic

Is extended lymphadenectomy necessary for surgical cure of gastric cancer?

No. Although results from numerous randomized trials have not shown an OS advantage of extended lymphadenectomy, CSS and LRR may be improved with extended dissection in the most recent update of the Dutch trial.

Dutch trial (Bonenkamp JJ et al., NEJM 1999): 711 pts randomized to D1 vs. D2 dissection. There was greater mortality in the D2 group (10% vs. 4%, p = SS), and 5-yr OS was 45% vs. 47% (NSS). In the most recent 15-yr update (Songun I et al., Lancet Oncol 2010), the 15-yr OS is 21% in the D1 group and 29% in the D2 groups (p = 0.34). However, the gastric cancer–related death rate is significantly higher in the D1 group (48%) vs. the D2 group (37%), while deaths from other Dz were similar in the 2 groups. LR was lower in the D2 group (12% vs. 22%) as well as the regional recurrence (13% vs. 19%) (all SS).

MRC trial (Cushieri A et al., Br J Cancer 1999): 400 pts randomized to D1 vs. D2. There was greater mortality in the D2 group (13% vs. 6.5%), and 5-yr OS was the same (35% vs. 33%).

In both trials, splenectomy and pancreatectomy had an adverse impact on survival.

Japanese trial JCOG9501 (D2 vs. D2 + P-A node dissection [PAND]) (Sasako M et al., NEJM 2008) demonstrated that although extended LND does not increase morbidity or mortality, there is also no difference in 5-yr OS (69.2% for D2 vs. 70.3% for D2 + PAND) or for LRR.

An Italian trial (D1 vs. D2) (Degiuli M et al., Br J Surg 2010) has shown no increased morbidity or mortality with extended lymphadenectomy, but clinical outcomes have not yet been reported.

What is the min # of LNs that should be pathologically assessed in a gastrectomy specimen?

In the U.S., at least 15 LNs should be assessed by the pathologist, since pt survival improves if >15 LNs are examined.

What are the selection criteria for endoscopic mucosal resection (EMR) or limited surgical resection (without nodal evaluation) of gastric cancer?

Favorable early-stage gastric cancer: Tis-T1 (but not involving submucosa), small (>3 cm), nonulcerated, well differentiated, N0. In general, these types of tumors have <5% LN met rate.

When is surgery alone adequate for gastric cancer?

T1N0 or T2N0 (but not beyond the muscularis propria). 5-yr OS for favorable early-stage gastric cancer is 80%–90%. For all others, adj Tx is needed (Ib [except T2aN0] to IIIC [nonmetastatic]).

What is the relapse pattern after “curative” resection of gastric cancer?

Distant Dz (50%) and LRR. LRR is common in the gastric bed, nearby LNs, anastomatic site, gastric remnant, and duodenal stump. In the classic paper of the University of Minnesota reoperative analysis (Gunderson L et al., IJROBP 1982), local-only recurrence was seen in 29%, LR and/or regional LN mets in 54%, and LF as any component of failure in 88% of pts.

What is the evidence that demonstrated the benefit of adj CRT after surgical resection for gastric cancer?

INT-0116 (Macdonald JS et al., NEJM 2001): 556 pts, stage IB–IV (nonmetastatic) adenocarcinoma of stomach and GE junction (~20%), randomized after en bloc resection to –margin to (a) observation or (b) CRT (1 cycle bolus 5-FU/leukovorin [LV]) before RT, 2 cycles during 45 Gy RT, and 2 cycles after RT. Median follow-up was 5 yrs. CRT was beneficial in all parameters except for DM 3-yr RFS 48% vs. 31%; 3-yr OS 50% vs. 41%; median OS 36 mos vs. 27 mos; and LR 19% vs. 29%. DM 18% surgery vs. 33% CRT (NSS). Toxic deaths of 1% were seen.

What is the major criticism for the benefit of CRT seen in INT-0116?

Suboptimal LND (54% D0, 10% D2) is the major criticism of INT-0116.

Is there a benefit of postop CRT for pts with more extensive lymphadenectomy?

Yes. There is a benefit according to a retrospective review from South Korea (Kim S et al., IJROBP 2005). In a series of 990 pts, stage II–IV with D2 resection +/− postop CRT, as done on INT-0116, postop CRT benefited all pts regardless of stage. Overall, CRT vs. surgery alone: 5-yr OS 57% vs. 51%, RFS 54.5% vs. 47.9%, and LRF 14.9% vs. 21%, respectively (all SS).

Is there a role for preop CRT for gastric cancer?

Possibly, although no phase III studies have been published to date. However, for GE junction tumors, preop CRT based on randomized data suggests benefit (Chapter 46: Walsh TN et al., NEJM 1996).

A phase II study of neoadj CRT (RTOG 9904: Ajani JA et al., JCO 2006) using induction chemo × 2 (5-FU/LV/cisplatin) → CRT (continuous infusion [CI] 5-FU/weekly Taxol) showed pCR of 26% and R0 resection of 77%.

Is there a role for postop RT alone for resected gastric cancer?

Possibly. Although not standard practice, there are weak data suggesting LC benefit with conflicting results on survival benefit. Most adj RT-alone trials utilized IORT +/− EBRT vs. surgery-alone randomization and found benefit with adj RT.

Is there data to show benefit of postop chemo alone after gastric cancer resection?

This is uncertain, with only 2 out of 6 randomized trials showing benefit. Most are small underpowered studies. The only good data with a large # of pts (Sakuramoto S et al., NEJM 2007) used adj S-1 (an oral fluoropyrimidine used in Japan) chemo vs. surgery alone in 529 pts, all with stage II or III gastric cancer with resection and D2 LND. S-1 had a superior 3-yr OS of 80% vs. 70% in the surgery-alone group (p = 0.002).

Recent meta-analysis of 3,658 pts from randomized trials showed a HR 0.82 with chemo. (Gianni L et al., Ann Oncol 2001)

What is the recent study that demonstrated a survival benefit of periop chemo compared to surgery alone for the management of gastric cancer? What is the major weakness with this approach?

MRC Adjuvant Gastric Cancer Infusional Chemo (MAGIC) trial (Cunningham D et al., NEJM 2006): 503 pts with gastric, GE junction, and distal esophageal adenocarcinoma (26%) randomized to (a) preop epirubicin/cisplatin/5-FU (ECF) × 3 and postop ECF × 3 or (b) surgery alone showed a survival benefit for chemo. 5-yr OS was 36% vs. 23%, respectively (p = 0.009). The major weakness of the MAGIC trial is that the pCR rate was 0%.

What is the current U.S. randomized trial in resected high-risk gastric cancer?

CALGB 80101, pts with completed resected, high-risk gastric cancer randomized to (a) the modified Macdonald regimen (5-FU/LV × 1 → CI 5-FU × 2 +RT → 2 cycles 5-FU/LV) or (b) ECF (epirubicin 50 mg/m2, cisplatin 60 mg/m2, and CI 5-FU 200 mg/m2/day × 21 days) × 1 → RT + CI 5-FU × 2 → ECF × 2 cycles.

What is the survival of pts with locally advanced unresectable gastric cancer? How are these pts managed?

5-yr OS is generally 5%–20%. In most randomized studies of these pts, CRT has benefit over chemo alone (5-yr OS 12%–18% vs. 0%–7%). GITSG G274 (Schein PS et al., Cancer 1982) used CRT (50 Gy) vs. chemo alone (5-FU/1-(2-chloroethyl)-3-(4-methylcyclohexyl)-1-nitrosourea [MeCCNU]). There was better survival with CRT (18% vs. 7%).

For metastatic gastric cancer pts, what are the palliative Tx for the incurable pt?

Surgical resection for carefully select pts with good performance status with Sx of obstruction or hemorrhage is better for palliation than stents or bypass. RT alone or CRT can be considered for the nonsurgical candidate (short course pallation 30 Gy/10 fx, 37.5 Gy/15 fx). Endoluminal laser ablation can be used for proximal lesions with esophageal obstruction → chemo. Palliative chemo compared to best supportive care has an overall HR of 0.39, and MS increases from 4.3 mos to 11 mos based on Cochrane meta-analysis (Wagner A et al., Cochrane Database Sys Rev 2006).

For the metastatic gastric cancer pt with bleeding or pain from the primary Dz, what are adequate palliative RT doses?

1.     Bleeding: 30 Gy in 10 fx may be sufficient; however, for the initial few doses, a higher fractional dose (4–4.5 Gy) may be better for bleeding control. After 3–4 fx, the practitioner can back down to 3 Gy/fx.

2.     Pain from tumor invasion: 45 Gy may be necessary.

What is the irradiation volume and dose of postop CRT after gastric tumor resection?

Tumor bed and nodal volumes constructed with preop/postop imaging and surgical clip placement. In general, node+ Dz requires wide coverage of the tumor bed, remaining stomach, all resection/anastomotic sites, and nodal drainage areas (which is dependent on tumor location). Use 45 Gy to the initial volume, then CD to 50.4 Gy to the surgical bed or at-risk areas. Pre-Tx J-tube placement (at time of staging laparoscopy) is helpful for nutritional support.

When would it be optional to treat the nodal beds in a resected gastric cancer?

Tx would be optional for pts with –nodes, pts having had adequate surgery and pathologic evaluation for nodes (>10–15 nodes), and wide surgical margins (at least 5 cm).

In general, what are the at-risk regional nodal sites based on anatomic location of the gastric tumor? Give the answer in terms of (1) GE junction, (2) proximal stomach, (3) body, and (4) distal stomach.

1.     GE junction: mediastinal, periesophageal, celiac, perigastric

2.     Proximal stomach: perigastric, periesophageal, celiac, pancreaticoduodenal, porta hepatis

3.     Body: perigastric, splenic, celiac, peripancreatic, porta hepatis

4.     Distal stomach: perigastric, periduodenal, peripancreatic, porta hepatis, celiac

How does the target volume differ for proximal vs. distal gastric lesions?

1.     Proximal lesions: include splenic hilum and left medial diaphragm in the target volume, but inf extent does not need to go to L3 (may just go to L1-2 coverage of the superior mesenteric artery/P-A nodes)

2.     Distal lesions: include 1st portion of the duodenal C-loop but not the splenic hilum

For proximal and distal lesions and –nodes with adequate dissection, the remnant of the stomach does not need to be covered. For body lesions, however, the gastric remnant needs to be covered in all cases.

What is the preferred technique for Tx planning of postop CRT?

3D-CRT (4 field) or IMRT can spare normal tissue and reduce toxicity better than traditional AP/PA fields, but greater conformality requires a more careful delineation of Tx targets.

image Toxicity

What are some long-term complications of gastrectomy?

Dumping syndrome (diarrhea, cramping, palpitations, reactive hypoglycemia) and malabsorption (B12, iron, calcium; supplement if necessary)

What is done during follow-up of pts treated with surgery and adj therapies?

Surgery and adj therapy follow-up: H&P every 4–6 mos × 3 yrs, then annually; CBC/CMP, endoscopy, and radiologic imaging as clinically indicated; and monitor for B12 deficiency (NCCN 2010)

In treating the postgastrectomy pt with adj CRT, what are the dose-limiting structures and dose limits for these organs?

Dose-limiting structures and dose limits:

1.     Remnant of stomach: 45 Gy

2.     Kidney: V18 <30%

3.     Liver: V30 <50%–70%

4.     Heart: If RT alone → V40 <50%, CRT for heart → V40 <40%

5.     Spinal cord: ≤45 Gy