Introduction
Esophageal cancer is the seventh most common cancer worldwide and the sixth leading cause of cancer-related death [
1], reflecting its generally poor prognosis. In 2018 in the United States, there were over 17,000 incident cases and 15,000 deaths [
2]. Overall 5-year survival is approximately 20% [
3], and most patients will die within one year of diagnosis [
4].
This situation creates a compelling need to minimize treatment-associated morbidity to provide esophageal cancer patients with the maximum quality of life for their often-limited expected survival. However, the primary treatment modality for esophageal cancer, surgical resection, has been associated with high morbidity and mortality. In a review of Medicare data between 1997 and 2003, esophagectomies (primarily open procedures) were associated with an inpatient mortality of 3.0% and a 30-day mortality of 14% [
5].
In an effort to improve on the outcomes of esophagectomy, Cuschieri, et al. developed and were the first to report on a minimally invasive esophagectomy (MIE) approach in 1992 [
6]. The technique was considered experimental until Luketich et al.’s landmark 2003 series of 222 MIE procedures, which were associated with an operative mortality of only 1.4% [
7]. Since that time, the procedure has been widely adopted and has progressively evolved as thoracic surgeons continue to incorporate additional innovations into the technique.
Despite the advantages over conventional open resections, there remain important areas for continued improvement in the process and postoperative outcomes associated with MIE procedures. Recent data show that median hospital length of stay (LOS) remains at approximately one week for MIE [
8‐
39]. Median intensive care unit (ICU) length of stay, when reported, is typically between 1 and 3 days for both open and MIE procedures [
9,
12,
14,
16,
17,
20,
26,
27,
30‐
32,
36,
37,
39,
40]. Reported 30-day readmission rates are typically 9–15% [
11,
18,
22,
32,
41,
42]. Although minimally invasive surgical approaches to esophagectomy have important advantages, reported complication rates remain in the range of 23–65% [
8,
10,
12,
13,
15‐
17,
19,
21‐
29,
29,
32,
33,
35,
36,
38‐
40,
42‐
46]. Furthermore, transabdominal jejunostomy feeding tubes are generally placed for post-operative nutrition, potentially causing postoperative complications [
47] and delaying return-to-baseline functional status; routine feeding tube placement is described in most protocols and in three large series, ≥ 84% of patients were discharged home with feeding tubes [
20,
37,
40]. Hence, there continue to be opportunities for further development and enhancement of the MIE approach. In addition, substantial advances have been made in perioperative care, such as enhanced-recovery-after-surgery (ERAS) protocols, providing potential additional improvements in patients' overall surgical experience.
Building on the notable progress made by prior surgical innovators, we sought to improve preoperative, intraoperative, and postoperative aspects of MIE, with the goal of improving esophageal cancer patients’ morbidity, mortality, and return-to-baseline functional status. We report here our experience with a streamlined fully minimally invasive operative approach along with a comprehensive perioperative care program.
Discussion
Esophageal cancer is a deadly disease, with an overall 5-year survival of only 19% [
4] and esophagectomy remains the mainstay of treatment. Since the introduction of the minimally invasive approach to esophagectomy in 1992 [
6], there have been important improvements in the technique and outcomes associated with the procedure, though many patients continue to be burdened by surgical morbidity, lifestyle restrictions (e.g., use of feeding tubes), and prolonged hospitalization. Accordingly, thoracic surgeons have continued to build on prior innovations to further improve on the promise of MIE [
60].
In this paper, we describe the process and clinical outcomes associated with a streamlined surgical approach to MIE developed at our institution among 142 consecutive patients with esophageal cancer treated with this procedure. This surgical technique was built around a fully minimally invasive approach incorporating updated techniques used by our group and others, motivated by prior advances in the procedure. In addition, we were careful to adhere to the principles of ERAS [
49] with careful patient preparation and close post-discharge outpatient follow-up, further optimizing both the surgical and perioperative experience. No patient was lost to follow-up for the assessment of 90-day postoperative outcomes.
In conjunction with well-coordinated preoperative and postoperative care, we found this approach was associated with excellent outcomes, substantially improving objective measures of surgical and perioperative performance as well as the patient experience. In particular, we found that this comprehensive perioperative and surgical approach was associated with relatively shorter operative times, decreased need for both intensive-care and overall inpatient hospitalization, reduced need for feeding tubes, and acceptable adverse-event, reoperation, and readmission rates.
For example, we observed a median inpatient LOS of 3 days in our series with no patient requiring a routine ICU admission and only 5.6% requiring a subsequent transfer to the ICU for management of complications. These results compare favorably with prior published MIE studies with reported hospital LOS's between 7 and 33 days [
8‐
40]; most studies reported LOS's of at least 12 days, with the largest case series reporting LOS’s of 8 to 15 days [
18,
20,
22,
32,
35]. The great majority of studies that provided data on intensive care unit LOS's reported a median of at least one day [
9,
12,
14,
16,
17,
20,
26,
27,
30‐
32,
36,
37,
39,
40,
61]. Furthermore, the elimination of a routine ICU admission and shorter overall LOS did not adversely affect rates of readmission or reoperation: our readmission rate was 12%, similar to previously reported rates of 9–18% [
11,
18,
22,
32,
41,
42]. In addition, only 19% (95% CI: 12.9% to 26.4%) of our patients required placement of a jejunostomy tube, compared with reported rates of 84% [
37], 95% [
20], and 97% [
40], among many other series that described routine jejunostomy-tube placement as part of the surgical protocol. We also found continued improvements comparing the earlier half of cohort with the latter half, suggesting that the learning curve for performing this procedure continued throughout the study period.
Comparing surgical complications across series is admittedly difficult since there is great variability in reporting standards. However, with respect to individual complications, we observed anastomotic leaks in 3 patients (2.1%, 95% CI: 0.4–6.0%) compared with others' reported rates that ranged from 0 to 21% [
8,
10‐
12,
14,
16,
17,
19,
21,
23‐
40,
42‐
44,
61‐
63] with a median rate of 10% and a range from 5.5 to 21% among the five largest series that reported relevant rates [
32,
35,
36,
42,
63]. In our series, pulmonary complications occurred in 14 patients (9.9%; 95% CI: 5.5% to 16.0%), including 9 cases (6.3%) of pneumonia and 2 cases (1.4%) of empyema; there were no cases of adult respiratory distress syndrome (ARDS) or chylothorax. These results compare favorably with rates from other centers with reported postoperative pneumonia rates ranging from 2 to 20% [
10,
13‐
16,
19,
22,
25,
28,
29,
35‐
38,
40,
43,
44] with the median reported rate of 8%, and reported rates of empyema ranging from 0 to 4.1% [
15,
20,
35,
37,
39,
42] with a median rate of 3.8%. Rates of ARDS reported in prior series range from 1 to 8% [
8,
10,
16,
20,
23,
35,
40,
43] with a median of 3.3%; reported rates of chylothorax range from 1 to 11% [
12,
14,
15,
17,
19,
23,
24,
26,
31,
32,
35,
37‐
40,
42,
44,
62] with the median reported rate of 3%. No patient died during the index hospitalization and 30-day mortality was 0.7% (95% CI: 0.02% to 3.9%) compared with reported rates in other series of 0–11% [
9,
11,
12,
16,
18‐
20,
22‐
24,
32,
34‐
37,
40,
43,
45,
61,
63] (with most in the range of 2–4%).
The reasons for the favorable outcomes we observed were likely multifactorial. While the fundamental surgical principles remained unchanged for the esophagectomy itself, we streamlined several technical aspects, eliminating the need for the pyloroplasty, jejunostomy and Kocherization. Meticulous attention was made to avoid tissue trauma to the gastric conduit by utilizing a “no-grab” technique, implementing partial conduit tubularization, and avoiding a linear gastrotomy. Having two subspecialty thoracic surgeons working together further reduced time under anesthesia.
The decrease in hospital LOS reflects not only improved intraoperative techniques, but also intensive perioperative management. Saving the right pleura and reconstituting the mediastinal envelope eliminated the need for a feeding tube and allowed early removal of the nasogastric tube with resultant initiation of oral nutrition on the first postoperative day. Of particular note, only 19% of our patients required placement of a feeding jejunostomy tube compared to much higher rates reported in the literature. Employing a closed suction drain provided continued chest drainage, which could be continued as an outpatient to monitor for leaks. Postoperative pain management was simplified by administering long-acting intercostal nerve blocks thereby eliminating epidural catheters. Early alimentation and ambulation allowed the patient to recover earlier at home. There was daily telephone communication with a staff surgeon and on-demand access as necessary to monitor patients' progress and address any patient concerns. Return visits were mostly limited to drain removal. The combination of a streamlined procedure, strict adherence to ERAS protocols by a highly coordinated perioperative team, and close postoperative follow-up all likely contributed to the favorable patient outcome and experience.
Our case series has several strengths, including a consecutive closed cohort, complete 90-day follow-up, and detailed clinical and utilization EMR data. Our patient sample was typical of patients in other cohorts and trials in terms of age, gender, ASA classification, cancer stage, comorbidities, and use of neoadjuvant therapy. However, several limitations of this report should be noted. First, this study was based on a single-center, retrospective design. Additionally, this was a study of a streamlined surgical approach along with instituting a centralized multidisciplinary care method; therefore, it was not possible to determine the impact of each individual component on the improved outcomes. Although several surgeons performed the new procedure, suggesting the results are not limited to the practice of a single practitioner, generalizability will need to be validated in other practice settings. Finally, the retrospective data collection did not allow for assessment of standardized quality-of-life assessments.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.