Back to 2017 Program and Abstracts

---

MULTIVISCERAL RESECTIONS FOR LOCALLY ADVANCED ESOPHAGO-GASTRIC CANCER: A BRIDGE TOO FAR?
Juan-Carlos Molina^*, Alreem Al Hinai, Alexandre Gosselin-Tardif, Carmen L. Mueller, Jonathan Spicer, David Mulder, Lorenzo E. Ferri
McGill University Health Centre, Montreal, QC, Canada

Introduction
Curative intent treatment for locally advanced esophagogastric cancer requires complete surgical resection with negative margins. For clinically T4b tumors with adhesion to adjacent organs, a multivisceral en-bloc resection is required to achieve this. However, the oncologic benefit of such major operations has been brought into question, particularly given a perceived unacceptable morbidity for patients with such high disease burden. Prior studies on this topic were hampered by historical data from an era in which chemotherapy was largely ineffective with low response rates. We thus sought to investigate the surgical and oncologic outcomes of en-bloc multivisceral resections for locally advanced gastric and esophagogastric junction (EGJ) carcinomas from a modern era cohort.
Material and Methods
A prospective North American institutional upper GI cancer database was queried from 2005 to 2015. We identified 515 patients who underwent resection for EGJ and gastric cancer, 140 of which had resections of organs in addition to the stomach and esophagus. For the purpose of this study, we included only patients who underwent curative intent en-bloc resection of adjacent organs for cT4b lesions. Clinicopathologic, perioperative and oncologic variables were assessed. Primary outcomes were perioperative complications, mortality, and overall survival. Cox proportional hazards model and multivariate logistic regression were employed. Data presented as Median(IQR).
Results
35 patients were included. Clinicopathologic and clinical data are presented in Table 1. The most common resected organs were pancreas (17, 49%), spleen (12, 34%) and liver (10, 29%). More than one organ other than stomach/esophagus were resected in 15(43%). Although all tumors were adherent to resected organs, in only 14 (42%), the final pathology confirmed organ invasion (pT4b). R0 resection rate was 94% (33). The majority (28, 80%) had lymph node involvement with a high nodal disease burden (positive/resected lymph nodes = 9(2-14)/33(19-47)). Postoperative complications occurred in 16(46%), 10 of which were Clavien-Dindo >2 (anastomotic leak (3, 9%) duodenal stump and pancreatic leak (3, 9%), bile leak, (2, 6%) enterocutaneous fistula (2, 6%)). LOS was 10 days (7-20) and 90-day mortality 0%. 1, 3, and 5-year survival was 82, 47 and 31%. On multivariate analysis, tumor size greater than 7 cm, lymph node involvement, and residual tumor after resection were independently associated with lower survival.
Conclusion
Multivisceral extended resections for esophagogastric cancer can achieve good long-term survival given the high burden of disease and is associated with an acceptable morbidity. R0 resection should be should remain the ultimate goal for the curative intent treatment for locally advanced gastric and EGJ carcinomas even if the tumor is adherent to adjacent organs.

Clinical and Pathologic outcomes

VARIABLE		N	%
Gender	male	10	29%
Age (Mean, IQR)		64.5 (59-75)
Location	EGJ	13	37
	Stomach	22	63
Surgery	Esophagectomy/proximal gastrectomy	11	31
	Total gastrectomy	14	40
	Extended total gastrectomy/esophagectomy	2	6
	Subtotal gastrectomy	8	23
N organs resected other than esophagus/stomach	Single	20	57
	Multiple	15	43
Organs resected	Pancreas	17
	Spleen	12
	Liver	10
	others (Colon, lung, adrenal)	18
pathologic T	T3	16	46
	T4a	5	14
	T4b	14	40
pathologic lymph nodes	N0	7	20
	N1-2	12	34
	N3	16	46
Postoperative outcomes	90-day mortality	0	0
	readmissions	4	11
	reoperations	6	17
Complications	No complication	19	54
	Clavien-Dindo I-II	6	17
	Clavien-Dindo IIIa	4	11
	Clavien-Dindo IIIb	4	11
	Clavien-Dindo IVa	2	6
	Total Complications	16	46

Table 2: Cox Univariate and Multivariate Regression Hazard Models

Feature		Cox Univariate Regression Model, HR (95% CI)	p value	Cox Multivariate Regression Model, HR (95% CI)	p value
Sex		1.4 (0.45-4.3)	0.55
Age >65		1.09 (0.41-2.9)	0.85
Multiple organs		1.58 (0.85-2.94)	0.14		0.3
	Pancreas	2.2 (0.80-6.04)	0.125		0.31
	Spleen	1.92 (0.72-5.14)	0.19
	Liver	0.93(0.30-2.9)	0.90
	Lung	0.85 (0.24-2.99)	0.79
	Other	1.13 (0.32-4.02)	0.85
Tumor Size (cm)		1.23 (1.04-1.45)	0.01		0.05
	Tumor size>7cm	2.52 (0.87-7.31)	0.09	3.28 (0.99-10.8)
Pathologic T stage		1.09 (0.65-1.84)	0.72
Lymph node involvement		11.5 (1.43-93.25)	0.02	39.4 (2.84-545.6)	0.006
Extent of resection (R1-2 vs 0)		1.77 (0.39-7.93)	0.45	6.6 (1.53-28.34)	0.011

---

Back to 2017 Program and Abstracts