Hydrogastric Sonography in thePreoperative Staging of Gastric Cancer

Jose Marıa Segura, PhD, Antonio Olveira, MD, Pilar Conde, MD, Jose Carlos Erdozain, PhD,Jose Suarez, MD

Gastroenterology Service, Hospital La Paz, Paseo de La Castellana, 261, 28041 Madrid, Spain

Received 5 January 1999; accepted 28 July 1999

ABSTRACT: Purpose. Depth of wall invasion is themain prognostic factor in gastric cancer. We studiedthe utility of hydrogastric sonography in the evalua-tion of transmural infiltration by gastric cancer.

Methods. Thirty-seven patients with gastric adeno-carcinoma were examined before surgery with a5-MHz probe after the ingestion of 100–400 ml of wa-ter (mean, 330 ml). Sonographic results were com-pared with pathologic classifications obtained aftersurgery.

Results. Of the 37 tumors, 15 were found at surgeryto be in the antrum, 10 were in the gastric body, 5were proximal, and 7 were diffuse. After surgery, tu-mors were classified as follows: 2 (5%) T1, 4 (11%) T2,15 (41%) T3, and 16 (43%) T4. Hydrogastric sonogra-phy correctly classified 30 (81%) of the 37 tumors. So-nography was correct for 2 (100%) of the 2 T1 tumors,2 (50%) of the 4 T2 tumors, 13 (87%) of the 15 T3tumors, and 13 (81%) of the 16 T4 tumors. Five sono-graphic errors were due to understaging and 4 tooverstaging. With regard to tumor site, sonographicresults were correct for 4 (57%) of the 7 diffuse tu-mors, 3 (60%) of the 5 proximal tumors, 9 (90%) of the10 gastric body tumors, and 14 (93%) of the 15 antraltumors.

Conclusions. Hydrogastric sonography is useful forpreoperative evaluation of transmural infiltration bygastric cancers, particularly tumors in the antrum orgastric body. © 1999 John Wiley & Sons, Inc. J ClinUltrasound 27:499–504, 1999.

Keywords: hydrogastric ultrasonography; gastric can-cer; cancer staging

Gastric cancer continues to be the second mostcommon cause of cancer-related deaths

worldwide despite the marked decrease in its in-

cidence in industrialized countries over the past60 years.1 About 22,800 cases are diagnosedyearly in the United States, with a mortality of14,700 per annum.2 Spain is among the nationsthat have an intermediate risk, with a standard-ized mortality rate of 13.82/100,000 inhabitants.3

Several prognostic factors have been established,the most important being the depth of gastric wallinvasion.4

Hydrogastric sonography (HGS) is a recentlydeveloped technique in which filling the stomachwith fluid allows satisfactory visualization of thelayers of the gastric wall and their disorders byconventional transabdominal sonography.5–8 Wereport our preliminary findings on the use of HGSin the preoperative evaluation of transmural in-filtration by gastric cancer.

PATIENTS AND METHODS

From September 1997 to September 1998, 46cases of gastric adenocarcinoma were diagnosedat our hospital by endoscopy and biopsy. Afterdiagnosis, patients underwent routine biochemi-cal studies, chest radiography, electrocardiogra-phy, and abdominal CT. Endoscopic sonographywas not performed. Six patients whose generalcondition was very poor or who had either distantmetastases or tumors that could not be resectedwere excluded from surgery, as were those whorefused surgery. Three additional patients, inwhom tumor was found in the fundus, were ex-cluded from the study because the tumors couldnot be seen on sonography. Thus, 37 patients [20men and 17 women; mean age, 66.2 years (range41–85 years)] were studied. All gave oral in-formed consent to participate.

Correspondence to: J. M. Segura

© 1999 John Wiley & Sons, Inc. CCC 0091-2751/99/090499-06

VOL. 27, NO. 9, NOVEMBER/DECEMBER 1999 499

All candidates for surgery underwent prospec-tive HGS during the week before the operation.The surgical team was blinded to the HGS resultsso that their decisions would not be influenced.Sonographic examinations were performed usingan EUB-415 ultrasound unit (Hitachi, Tokyo, Ja-pan) with a 5-MHz transducer. Scanning wasdone in transverse, longitudinal, and obliqueplanes after the ingestion of drinking water(mean, 330 ml; range, 100–400 ml). The patientwas scanned first in the supine position and thenin the right or left lateral decubitus or Trendelen-burg position, as required to best visualize thelesion. All scanning was done by 1 person (JMS),who has 18 years of experience in sonographictechniques. That person knew the location of thetumor from previous endoscopy but was unawareof the results of other studies that could demon-strate the size or extent of the tumor infiltration.The average examination time was 15 minutes,and no complications were recorded.

The layers of the normal gastric wall wereidentified according to previous descriptions9–11

as follows: the innermost hyperechoic layer corre-sponds to the superficial mucosa; the second layeris hypoechoic and corresponds to the deep mucosa;the third layer is hyperechoic and corresponds tothe submucosa plus the acoustic interface be-tween the submucosa and muscularis propria; thefourth layer, which is hypoechoic, corresponds tothe muscularis propria minus the acoustic inter-face between the submucosa and the muscularispropria; and the fifth layer, which is hyperechoic,corresponds to the subserosal fat and serosa.

On HGS, stomach cancer was identified as ahypoechoic irregular mass that altered the sono-

graphic structure of the gastric wall. Lymph nodemetastases were not evaluated by HGS. In accor-dance with Caletti et al,12 the neoplasms wereclassified sonographically as follows: T1 (Figure1), abnormal findings limited sonographically tothe third layer, with the fourth layer remainingintact; T2 (Figure 2), complete interruption ofthe central layer with invasion of the fourth layer;T3 (Figure 3), interruption of the outermost hy-perechoic layer; and T4 (Figure 4), invasion ofadjacent organs. These assessments were com-pared with pathologic staging done after sur-gery.13

RESULTS

Of the 37 tumors studied, 15 were found at sur-gery to be in the antrum, 10 in the gastric body,and 5 in the fundus or subcardiac regions; 7 werediffuse. Pathologic staging was as follows: 2 tu-mors (5%) were classified as T1, 4 (11%) as T2, 15(41%) as T3, and 16 (43%) as T4. Five of the T4tumors were classified as such at laparotomy (3had invaded the pancreas and 1 the liver, andanother was disseminated through the perito-neum) and were not removed at surgery.

HGS correctly classified 30 (81%) of the 37 tu-mors. Sonographic assessment of wall infiltrationwas correct for 2 (100%) of the 2 T1 tumors,2 (50%) of the 4 T2 tumors, 13 (87%) of the15 T3 tumors, and 13 (81%) of the 16 T4 tumors.Five errors were due to understaging: 2 T3 tu-mors were classified as T1 and T2 and 3 T4 tu-mors as T3. These latter 3 tumors had invadedthe mesocolon and pancreas, greater omentum,

FIGURE 1. Transverse sonogram of a fluid-filled gastric lumen (GL) showing a T1 antral tumor (arrowheads)measuring 11 × 26 mm (D1).

SEGURA ET AL

500 JOURNAL OF CLINICAL ULTRASOUND

and peritoneum. Two errors were due to overstag-ing: 2 T2 tumors were classified as T3. With re-gard to tumor site, sonography correctly identi-fied 4 (57%) of the 7 diffuse tumors, 3 (60%) of the5 proximal tumors, 9 (90%) of the 10 tumors in thegastric body, and 14 (93%) of the 15 antral tu-mors.

DISCUSSION

The treatment of choice for gastric cancer is stillsurgical resection,14 although new treatmentstrategies are being studied. In early cancer, local

treatment by endoscopic resection with neoadju-vant chemotherapy produces satisfactory re-sults.15–18 Neoadjuvant chemotherapy also seemsto improve the overall survival rate of patientswith inoperable tumors.19,20 With regard to sur-gery, particular interest is focused on decreasingthe morbidity and cost associated with nonthera-peutic laparotomy for abdominal metastases un-detected by CT21; such cases continue to repre-sent up to 26% of the laparotomies conducted forgastric cancer.22

Choosing the best treatment for each patientrequires diagnosing the exact depth of tumor pen-

FIGURE 2. (A) Transverse sonogram showing a normal 5-layer gastric wall measuring 4.4 mm thick (calipers,D1). (B) Anterior wall T2 tumor (arrow) measuring 7.1 mm (calipers, D2).

FIGURE 3. Midline longitudinal sonogram of a fluid-filled gastric lumen (GL) showing a T3 tumor (arrow)without invasion of the liver (L).

HYDROGASTRIC SONOGRAPHY FOR GASTRIC CANCER

VOL. 27, NO. 9, NOVEMBER/DECEMBER 1999 501

etration. Although endoscopy and double-contrastradiography have significantly improved the ac-curacy of diagnosis of gastric cancer, neither pro-vides information on the depth of tumor invasion.The staging method used most often at present isCT. However, the accuracy of dynamic CT for Tstaging over all stages is only 65%, despite an83% accuracy rate for detecting serosal invasionand therefore advanced disease.23 Moreover, thesensitivity of modern spiral CT is only 50%, 71%,and 76% for detecting pancreatic invasion, peri-toneal metastases, and invasion of the colon ormesocolon, respectively.24 More sophisticatedstaging techniques such as MRI,25 endoscopicMRI,26 and laparoscopic sonography27 have pro-duced 88%, 89%, and 92% accuracy rates, respec-tively, at T staging. Endoscopic sonography hasbeen the most widely studied of the new stagingmethods; the accuracy of this technique hasranged from 67%28 to 92%,29 with a mean accu-racy of 77%.30

Our results compare favorably with those gen-erated with these new techniques, not only interms of diagnostic accuracy (81%) but also withregard to cost and noninvasiveness. However, ourresults are only preliminary. The low number ofpatients with early-stage cancer reflects not pre-selection of patients but rather the limited num-bers (< 10%) of tumors diagnosed at early stagesin the Western countries.31 Additional patientswith early-stage cancer are necessary to obtainadequate data for this subgroup. Although someJapanese studies have focused on staging gastriccancer in its early stages,35,36 interest in the Westhas been directed more toward preoperative as-

sessments of the operability of late-stage tu-mors,27,32–34 particularly since nonsurgical pallia-tive measures such as stenting and endoscopiclaser therapy have become available. Althoughour patients were not evaluated for operability,the accuracy of sonography for T3 and T4 tumorsin our study (87% and 81%, respectively) com-pares favorably with the rates obtained by endo-scopic sonography (91% for T3 and 80% for T4tumors).37 We believe that these T3 and T4 stag-ing results predict that HGS will be accurate inpreoperative evaluations of resectability and thatfurther study is warranted.

In our study, the 3 instances in which T4 tu-mors were identified sonographically as T3 tu-mors involved invasion of the pancreas, mesoco-lon, peritoneum, and omentum, sites that arevisualized with difficulty by other techniquessuch as CT and endoscopic sonography.21,24,29 T3understaging and T2 overstaging could also beexpected based on previous endoscopic sonogra-phy reports.30

Our results were better for distal tumors (ie,those in the gastric body or antrum) than forproximal or diffuse tumors. This finding, whichcould be anticipated,7 probably reflects the prox-imity of the antrum to the skin and the tendencyof gas to accumulate in the fundus. We obtainedgood views in 37 (93%) of 40 patients.

We found only 1 published study on staginggastric cancer by HGS.38 The results of those au-thors were similar to ours, with an 86% accuracyrate for diagnosing tumors as early or advanced.However, the authors did not classify the tumorsaccording to T stage.

FIGURE 4. Transverse sonogram of a T4 antral tumor (arrowheads) with posterior invasion of the retroperi-toneal space (arrow). GL, gastric lumen.

SEGURA ET AL

502 JOURNAL OF CLINICAL ULTRASOUND

In conclusion, HGS is useful in the preopera-tive evaluation of the degree of transmural infil-tration by gastric cancers, particularly for tumorsof the gastric body or antrum. Further studies ofthis technique are needed to evaluate its ability topredict the operability of stomach cancer. HGScould also be valuable for reducing the number oflaparotomies for nonresectable lesions, especiallyin areas where access to sophisticated imagingtechniques is limited.

REFERENCES

1. Fuchs CS, Mayer RJ. Gastric carcinoma. N Engl JMed 1995;333:32.

2. Neugut AI, Hayek M, Howe G. Epidemiology ofgastric cancer. Semin Oncol 1996;23:281.

3. Cortina P, Gil A, Saiz C, et al. Evolucion de losniveles de refrigeracion domestica de alimentos enEspana y mortalidad por cancer gastrico. Rev EspEnf Dig 1988;74:435.

4. Davis GR. Neoplasm of the stomach. In: SleisengerMH, Fordtran JS, editors. Gastrointestinal dis-ease. 5th ed. Philadelphia: WB Saunders; 1993. p.763–789.

5. Worlicek H, Lederer P, Lux G. Ultrasonographicevaluation of the wall of the fluid-filled stomach—case report of a leiomyoblastoma. Hepatogastroen-terology 1986;33:184.

6. Worlicek H, Dunz D, Engelhard K. Ultrasonic ex-amination of the wall of the fluid-filled stomach. JClin Ultrasound 1989;17:5.

7. Joharjy IA, Mustafa MA, Zaidi AJ. Fluid-aided so-nography of the stomach and duodenum in the di-agnosis of peptic ulcer disease in adult patients. JUltrasound Med 1990;9:77.

8. Yeh HC, Rabinowitz JG. Ultrasonography andcomputed tomography of gastric wall lesions. Ra-diology 1981;141:147.

9. Aibe T, Fuji T, Okita K, et al. A fundamental studyof normal layer structure of the gastrointestinalwall visualized by endoscopic ultrasonography.Scand J Gastroenterol 1986;21(suppl 123):6.

10. Bolondi L, Caletti GC, Casanova P, et al. Problemsand variations in the interpretation of the ultra-sound feature of the normal upper and lower gas-trointestinal tract wall. Scand J Gastroenterol1986;21:16.

11. Kimmey MB, Martin RW, Haggitt RC, et al. His-tologic correlates of gastrointestinal ultrasoundimages. Gastroenterology 1989;96:433.

12. Caletti GC, Ferrari A, Brocchi E, et al. Accuracy ofendoscopic ultrasonography in the diagnosis andstaging of gastric cancer and lymphoma. Surgery1993;113:14.

13. American Joint Committee on Cancer manual forstaging of cancer. 4th ed. Philadelphia: JB Lippin-cott; 1993.

14. Brennan MF, Karpeh MS. Surgery for gastric can-cer: the American view. Semin Oncol 1996;23:352.

15. Sibille A, Descamps C, Jonard P, et al. EndoscopicNd:YAG treatment of superficial gastric carci-noma: experience in 18 Western inoperable pa-tients. Gastrointest Endosc 1995;42:340.

16. Otsuka S. Endoscopic treatment of early gastriccancer with laser and laserthermia using Nd:YAGlaser. Nippon Rinsho 1996;54:1326.

17. Otsuka S. Endoscopic treatment of early gastriccancer with heat probe method. Nippon Rinsho1996;54:1351.

18. Kumagai K. The significance of preoperative che-motherapy for early gastric carcinoma. Surg Today1993;23:875.

19. Yonemura Y, Sawa T, Kinoshita K, et al. Neoadju-vant chemotherapy for high-grade advanced gas-tric cancer. World J Surg 1993;17:256.

20. Shinoda M, Morise K, Kusugami K, et al. Combi-nation chemotherapy with FP versus FEP in pa-tients with advanced gastric cancer. Researchgroup of gastric cancer chemotherapy. Gan ToKagaku Ryoho 1995;22:515.

21. Lowy AM, Mansfield PF. Gastric carcinoma [let-ter]. N Engl J Med 1995;333:1426.

22. Cady B, Ramsden DA, Stein A, et al. Gastric can-cer. Contemporary aspects. Am J Surg 1997;133:423.

23. Cho JS, Kim JK, Rho SM, et al. Preoperative as-sessment of gastric carcinoma: value of two-phasedynamic CT with mechanical IV injection of con-trast material. Am J Roentgenol 1994;163:69.

24. Davies J, Chalmers AG, Sue-Ling HM, et al. Spiralcomputed tomography and operative staging ofgastric carcinoma: a comparison with histopatho-logical staging. Gut 1997;41:314.

25. Matsushita M, Hiromichi O, Murakami T, et al.Extraserosal invasion in advanced gastric cancer:evaluation with MR imaging. Radiology 1994;192:87.

26. Inui K, Nakazawa S, Yoshino J, et al. EndosocpicMRI: preliminary results of a new technique forvisualization and staging of gastrointestinal tu-mors. Endoscopy 1995;27:480.

27. Finch MD, John TG, Garden OJ, et al. Laparoscop-ic ultrasonography for staging gastroesophagealcancer. Surgery 1997;121:10.

28. Ohashi S, Nakazawa S, Yoshino J. Endoscopic ul-trasonography in the assessment of invasive gas-tric cancer. Scand J Gastroenterol 1989;24:1039.

29. Botet JF, Lightdale CJ, Zauber AG, et al. Preop-erative staging of gastric cancer: comparison of en-doscopic US and dynamic CT. Radiology 1991;181:426.

30. Rosch T. Endosonographic staging of gastric can-cer: a review of literature results. Gastrointest En-dosc Clin North Am 1995;5:549.

31. Wanebo HJ, Kennedy BJ, Chmiel J, et al. Cancer ofthe stomach: a patient care study by the AmericanCollege of Surgeons. Ann Surg 1993;218:583.

32. Tio TL, den Hartog Jager FCA, Tytgat GNJ. The

HYDROGASTRIC SONOGRAPHY FOR GASTRIC CANCER

VOL. 27, NO. 9, NOVEMBER/DECEMBER 1999 503

role of endoscopic ultrasonography in assessing lo-cal resectability of oesophagogastric malignancies.Accuracy, pitfalls and predictability. Scand J Gas-troenterol 1986;21(suppl 123):78.

33. Rosch T, Lorenz R, Zenker K, et al. Local stagingand assessment of resectability in carcinoma of theesophagus, stomach, and duodenum by endoscopicultrasonography. Gastrointest Endosc 1992;38:460.

34. Mortensen MB, Pedersen SA, Hovendal CP. Preop-erative assessment of resectability in gastroesoph-ageal carcinoma by linear array endoscopic ultra-sonography. Scand J Gastroenterol 1994;29:341.

35. Abe S, Lightdale CJ, Brennan MF. The Japaneseexperience with endoscopic ultrasonography in thestaging of gastric cancer. Gastrointest Endosc1993;39:586.

36. Yanai H, Tada M, Karita M, et al. Diagnostic util-ity of 20-megahertz linear endoscopic ultrasonog-raphy in early gastric cancer. Gastrointest Endosc1996;44:29.

37. Pollack BJ, Chak A, Sivak MV. Endoscopic ultra-sonography. Semin Oncol 1996;23:336.

38. Lim JH, Ko YT, Lee DH. Transabdominal US stag-ing of gastric cancer. Abdom Imaging 1994;19:527.

SEGURA ET AL

504 JOURNAL OF CLINICAL ULTRASOUND