Introduction: Inadvertent thoracic duct injury is common during esophagectomy and may result in postoperative chylothorax. This study’s objective was to investigate utility of patent blue injection as a modality for intraoperative thoracic duct visualization. Methods: A prospective, single-arm, interventional study of patients undergoing minimally invasive esophagectomy was performed. Patients were injected with patent blue dye into both groins prior to thoracic stage of surgery and assessed for duct visualization. Control group was formed by propensity score matching using retrospectively collected data regarding patients who underwent esophagectomy. Results: A total of 25 patients were included in analysis, compared to a control of 50 patients after matching. Thoracic duct was visualized in 60% of patients in the study group (15/25 patients). Significant differences were found between study and control groups (p < 0.05) with regards to median operative time (422 vs. 285 min, respectively), overall complications (16 vs. 34%, respectively), and median postoperative length of stay (13.5 vs. 10 days, respectively). There was a difference in rate of chyle leak between study and control groups; however, this was not significant (0 vs. 12%, respectively, p = 0.17). Conclusion: Patent blue injection represents a simple method for thoracic duct visualization during minimally invasive esophagectomy which may improve surgical outcomes.

Esophagectomy, with or without perioperative adjunctive therapy, is the main treatment option for locoregional esophageal cancer [1]. Esophagectomy is a major procedure with high incidence of perioperative morbidity and mortality [2]. Inadvertent thoracic duct injury, which occurs in 2–12% of patients [3], may result in chylothorax, ending in mortality in up to 18% of patients [4]. Several reports addressed risk factors for postoperative chylothorax, including squamous cell pathology [5, 6], low BMI (<25 kg/m2) [4, 7], low preoperative albumin levels [8], neoadjuvant therapy [3, 6, 9], the operating surgeon’s experience [10], and number of harvested lymph nodes [8]. Several studies addressed treatment options for postoperative chylothorax [4, 11‒19]. However, only few have investigated preventive interventions and measures [20‒23] or thoracic duct identification [24‒29].

Study Endpoints

The primary endpoint of this study was intraoperative thoracic duct visualization. Secondary endpoints include prevention of postoperative chyle leak and documentation of other types of postoperative morbidity.

Study Design and Patient Population

A prospective study with a single intervention arm was performed. All patients over 18 years of age candidates for minimally invasive esophagectomy (MIE) were offered the option of participation in this study. Patients with known allergic response to patent blue, pregnant women, and open procedures were excluded from the study. All methods were carried out in accordance with relevant guidelines and regulations, and informed consent was obtained from all subjects.

Data pertaining to demographics, comorbidities, and use of neoadjuvant therapy were obtained. Operative data were recorded, including surgical approach, operative time, intraoperative use of blood products, anastomotic techniques, and thoracic duct visualization. Postoperative outcomes were evaluated including complications according to Clavien-Dindo classification system [30] and escalation of care (defined as clinical deterioration requiring admission to the intensive care unit), final pathological reports, and 30- and 90-day mortality.

Surgical and Perioperative Variables

Patients were subjected to either minimally invasive three field or Ivor-Lewis esophagectomy depending on tumor’s location, and patients were subcutaneously injected with 2 mL of 2.5% patent blue V sodium (GuerbetR) into both groins immediately before the thoracic phase of the procedure. Patent blue concentration was chosen based on previously described methods for lymph node mapping utilized in breast cancer surgery [31]. Moreover, higher concentrations have previously been suggested to carry increased risks of adverse reactions [32]. Following consultation with the medical center’s angiographic team and following literature review of previously described injection sites for thoracic duct visualization, including ultrasound guided injection directly into the groin lymph nodes, injecting into the foot dorsum, and bilateral superficial groin injection [33]; it was determined that bilateral groin injection would offer high reproducibility and simplicity, with the least need for operator proficiency or additional imaging devices. For MITF esophagectomy, patient is placed in a left lateral decubitus position, with one lung ventilation. Three 12-mm trocars are then introduced in the following locations: 7th intercostal space at the posterior axillary line, 9th intercostal space at the mid-axillary line, and 4th intercostal space at the anterior axillary line. The pleural cavity is insufflated with CO2 up to 6-mm Hg pressure. The inferior pulmonary ligament is transected, and pleura is opened to reveal the esophagus. The esophagus is mobilized at its entire length from the hiatus to the azygous vein. The azygos vein is transected with a vascular endo-stapler (laparoscopic Vascular EndoGIA EthiconR). Lymphatic tissue is dissected along the esophagus according to the corresponding lymph node stations. Duct visualization during the thoracic phase is assessed with approximately 10–15 min from injection of patent blue. Duct assessment is performed according to anatomical landmarks from the hiatus up to the thoracic inlet.

Statistical Analysis

Data analysis was performed using SPSS version 23.0. Mann-Whitney (Wilcoxon) test was used for intergroup comparisons for continuous variables, while the χ2 test or Fisher’s exact test were used to compare categorical data. Data collected retrospectively regarding patients who underwent esophagectomy without patent blue injection was used as control for this study. In order to reduce effects of potential confounding factors in the comparison between two groups, a propensity score (PS) [34] was calculated to create well-balanced groups using R Statistical Software. PS was estimated using a multivariate logistic regression model with the treatment groups as the dependent variables and potential confounders as covariates. All patients in the patent blue group (PBG) were matched 1:2 to the control group according to PS using the global optimum method.

PS matching was performed according to age, sex, smoking status, alcohol use, existence of any other comorbidities aside from esophageal tumor (including ischemic heart disease, chronic heart failure, diabetes mellitus, hypertension, chronic obstructive pulmonary disease, chronic kidney disease, and atrial fibrillation), tumor location (according to Siewert-Stein classification [35]), whether patient received neoadjuvant treatment, year when the operation was performed, and type of procedure performed (Ivor-Lewis/Orringer/McKeown) and squamous pathology.

Twenty-six patients consented for the study and were initially eligible for the interventional arm. One patient was eventually excluded from data analysis since he underwent a trans-hiatal esophagectomy (the thorax was not accessed and dye was not assessed), resulting in 25 patients in the PBG, who served as the study group. A total of 220 patients who underwent esophagectomy during the same period were included in a preliminary control group prior to the PS matching. The final control group included 50 patients after PS matching.

Patent blue injection has previously been described to result in both local and systemic side effects, including allergic reactions of various degrees, skin tattooing at injection site, and a general bluish skin hue that may persist for up to 2 days from the time of injection [36, 37]. As patients with known allergy to patent blue were primarily excluded from this study, aside from a self-limiting skin tattooing at injection sites, no side effects related to dye injection were recorded.

Demographic data are presented in Table 1. Median age in the study group was 68 years (IQR 57–72) with a male predominance of 21/25 patients (84%). No differences were demonstrated in median preoperative albumin, neither before nor after propensity score matching (p > 0.05). Forty percent of patients (10 patients) were positive for smoking status, and 1 patient was positive for history of alcohol use. Adenocarcinomas were found in 76% (19 patients), and squamous cell carcinomas were found in 20% (5 patients). More patients in the study group received FLOT neoadjuvant therapy compared to control after PS matching (33.3 vs. 6%, p = 0.004). However, no differences were found in overall utilization of neoadjuvant radiation between groups (60 vs. 56%, p = 0.087).

Table 1.

Demographic and preoperative variables

Before matchingAfter matching
Variablesno patent (n = 227)patent (n = 25)p valueno patent (n = 50)patent (n = 25)p value
Median age, years (IQR) 67 (60–73) 68 (57–72) 0.983 65.5 (60–70) 68 (57–72) 0.408 
Male sex, n (%) 150 (66) 21 (84) 0.069 33 (66) 21 (81) 0.102 
Median BMI, kg/m2 (IQR) 24.85 (22.1–28.3) 25.9 (22.1–25.9) 0.887 23.4 (20.8–26.7) 25.9 (22.1–25.9) 0.320 
Median preop albumin (IQR) 4 (3.7–4.3) 4.05 (3.8–4.4) 0.364 4 (3.7–4.4) 4.1 (3.8–4.4) 0.995 
Smoker, n (%) 33 (15) 10 (40) 0.012 10 (20) 10 (40) 0.065 
Alcohol use, n (%) 3 (1.5) 1 (4) 0.375 1 (2) 1 (4) >0.99 
Comorbidity, n (%) 138 (60.8) 11 (44) 0.105 32 (64) 11 (44) 0.099 
Tumor type, n (%) 
 Adenocarcinoma 165 (73.3) 19 (76) 0.774 37 (74) 19 (76) 0.851 
 Squamous 60 (27.6) 5 (20) 0.471 13 (26) 5 (20) 0.566 
 NET 0 (0) 2 (8) 0.010 0 (0) 3 (8) 0.108 
Tumor location, n (%) 
 GEJ 100 (44.1) 18 (72) 0.024 25 (50) 18 (72) 0.069 
 Lower esophagus 75 (36.1) 6 (24) 0.232 16 (32) 6 (24) 0.473 
 Mid-esophageal 30 (14.4) 1 (4) 0.215 7 (14) 1 (4) 0.256 
 Upper esophagus 3 (1.4) 0 (0) >0.99 2 (4) 0 (0) 0.550 
Received neoadjuvant, n (%) 163 (71.8) 22 (88) 0.082 43 (86) 22 (88) >0.99 
Neoadjuvant therapy, n (%) 
 CROSS 96 (43.8) 11 (45.8) >0.99 26 (52) 24 (45.8) 0.619 
 FLOT 3 (1.4) 8 (33.3) <0.001 3 (6) 8 (33.3) 0.004 
 Experimental protocol 26 (11.9) 0 (0) >0.99 1 (2) 0 (0) >0.99 
 Other therapy 48 (21) 1 (4) 0.0398 5 (10) 1 (4) 0.815 
Neoadjuvant radiation, n (%) 88 (38.8) 14 (56) 0.115 30 (60) 14 (56) 0.087 
Before matchingAfter matching
Variablesno patent (n = 227)patent (n = 25)p valueno patent (n = 50)patent (n = 25)p value
Median age, years (IQR) 67 (60–73) 68 (57–72) 0.983 65.5 (60–70) 68 (57–72) 0.408 
Male sex, n (%) 150 (66) 21 (84) 0.069 33 (66) 21 (81) 0.102 
Median BMI, kg/m2 (IQR) 24.85 (22.1–28.3) 25.9 (22.1–25.9) 0.887 23.4 (20.8–26.7) 25.9 (22.1–25.9) 0.320 
Median preop albumin (IQR) 4 (3.7–4.3) 4.05 (3.8–4.4) 0.364 4 (3.7–4.4) 4.1 (3.8–4.4) 0.995 
Smoker, n (%) 33 (15) 10 (40) 0.012 10 (20) 10 (40) 0.065 
Alcohol use, n (%) 3 (1.5) 1 (4) 0.375 1 (2) 1 (4) >0.99 
Comorbidity, n (%) 138 (60.8) 11 (44) 0.105 32 (64) 11 (44) 0.099 
Tumor type, n (%) 
 Adenocarcinoma 165 (73.3) 19 (76) 0.774 37 (74) 19 (76) 0.851 
 Squamous 60 (27.6) 5 (20) 0.471 13 (26) 5 (20) 0.566 
 NET 0 (0) 2 (8) 0.010 0 (0) 3 (8) 0.108 
Tumor location, n (%) 
 GEJ 100 (44.1) 18 (72) 0.024 25 (50) 18 (72) 0.069 
 Lower esophagus 75 (36.1) 6 (24) 0.232 16 (32) 6 (24) 0.473 
 Mid-esophageal 30 (14.4) 1 (4) 0.215 7 (14) 1 (4) 0.256 
 Upper esophagus 3 (1.4) 0 (0) >0.99 2 (4) 0 (0) 0.550 
Received neoadjuvant, n (%) 163 (71.8) 22 (88) 0.082 43 (86) 22 (88) >0.99 
Neoadjuvant therapy, n (%) 
 CROSS 96 (43.8) 11 (45.8) >0.99 26 (52) 24 (45.8) 0.619 
 FLOT 3 (1.4) 8 (33.3) <0.001 3 (6) 8 (33.3) 0.004 
 Experimental protocol 26 (11.9) 0 (0) >0.99 1 (2) 0 (0) >0.99 
 Other therapy 48 (21) 1 (4) 0.0398 5 (10) 1 (4) 0.815 
Neoadjuvant radiation, n (%) 88 (38.8) 14 (56) 0.115 30 (60) 14 (56) 0.087 

IQR, interquartile range; GEJ, gastroesophageal junction; other therapy, cisplatin+5fu, ECX, or FOLFOX.

Tumor size, lymph node involvement, lymphovascular, or perineural invasion were similar in both groups (Table 2). A difference was found for median number of harvested lymph nodes before PS matching (19 nodes (IQR 13–29.5) with use of patent blue versus 13 nodes (IQR 9–18) without, p < 0.001]. This difference, however, was not significant after matching (18 nodes [12–23] without patent blue, p = 0.366).

Table 2.

Pathological variables

Before matchingAfter matching
Variablesno patent (n = 227)patent (n = 25)p valueno patent (n = 50)patent (n = 25)p value
Median tumor size, cm (IQR) 2.5 (1.5–3.5) 2.5 (1.8–4) 0.524 2.6 (1.7–4) 2.5 (1.8–4) 0.981 
pT, n (%)   0.089   0.652 
 0 49 (21.8) 4 (19)  12 (24) 4 (19)  
 1 44 (19.6) 2 (9.5)  4 (8) 2 (9.5)  
 2 37 (16.4) 6 (28.6)  12 (24) 6 (28.6)  
 3 95 (42.2) 8 (38.1)  22 (44) 8 (38.1)  
 4 0 (0) 1 (4.8)  0 (0) 1 (4.8)  
Median number of LN harvested (IQR) 13 (9–18) 19 (13–29.5) <0.001 18 (12–23) 19 (13–29.5) 0.366 
Median number of positive LN (IQR) 0 (0–1) 1 (0–3) 0.099 0 (0–2.5) 1 (0–3) 0.102 
LVI, n (%) 33 (20.4) 4 (16.7) 0.790 8 (21.1) 4 (16.7) 0.752 
PNI, n (%) 42 (25.8) 6 (25) >0.99 13 (34.2) 6 (25) 0.574 
Before matchingAfter matching
Variablesno patent (n = 227)patent (n = 25)p valueno patent (n = 50)patent (n = 25)p value
Median tumor size, cm (IQR) 2.5 (1.5–3.5) 2.5 (1.8–4) 0.524 2.6 (1.7–4) 2.5 (1.8–4) 0.981 
pT, n (%)   0.089   0.652 
 0 49 (21.8) 4 (19)  12 (24) 4 (19)  
 1 44 (19.6) 2 (9.5)  4 (8) 2 (9.5)  
 2 37 (16.4) 6 (28.6)  12 (24) 6 (28.6)  
 3 95 (42.2) 8 (38.1)  22 (44) 8 (38.1)  
 4 0 (0) 1 (4.8)  0 (0) 1 (4.8)  
Median number of LN harvested (IQR) 13 (9–18) 19 (13–29.5) <0.001 18 (12–23) 19 (13–29.5) 0.366 
Median number of positive LN (IQR) 0 (0–1) 1 (0–3) 0.099 0 (0–2.5) 1 (0–3) 0.102 
LVI, n (%) 33 (20.4) 4 (16.7) 0.790 8 (21.1) 4 (16.7) 0.752 
PNI, n (%) 42 (25.8) 6 (25) >0.99 13 (34.2) 6 (25) 0.574 

IQR, interquartile range; LN, lymph nodes; LVI, lymphovascular invasion; PNI, perineural invasion.

The thoracic duct was visualized in 15 of 25 patients in the study group (60%; Fig. 1). When visible, dye was found in the areas of the carina and along the trachea, and less frequently, it was seen cephalad to the diaphragm. Significant differences were found between groups after PS matching both in surgeons performing the procedure (Table 3) and in median operative time, which was longer for the study group (422 min [IQR 278–531] versus 285 min [IQR 253–335], p = 0.003). No significant differences were found in surgical approach or use of hand sewn versus stapled anastomosis. Moreover, no single surgeon was associated with the occurrence of chyle leak over another (p > 0.05).

Fig. 1.

a, b Visualization of the thoracic duct following patent blue injection: duct can be seen tinted blue, between dashed lines.

Fig. 1.

a, b Visualization of the thoracic duct following patent blue injection: duct can be seen tinted blue, between dashed lines.

Close modal
Table 3.

Operative and postoperative variables

Before matchingAfter matching
Variablesno patent (n = 227)patent (n = 25)p valueno patent (n = 50)patent (n = 25)p value
Surgical 
 Surgery type, n (%) 
  Trans-hiatal esophagectomy 45 (19.9) 0 (0) 0.011 2 (4) 0 (0) 0.550 
  Transthoracic esophagectomy 23 (10.2) 10 (40) <0.001 17 (34) 10 (40) 0.610 
  3-field esophagectomy 158 (69.9) 15 (60) 0.310 31 (62) 15 (60) >0.99 
 Surgery approach, n (%) 
  Laparoscopic/Robotic 179 (79.2) 22 (91.7) 0.182 48 (96) 23 (92) >0.99 
  Open 19 (43) 1 (4.2) 0.089 2 (4) 1 (4.2) >0.99 
 Surgeons, n (%) 
  1 174 (85.37) 10 (40) <0.001 34 (68) 10 (40) 0.020 
  2 136 (66.7) 13 (52) 0.147 45 (90) 13 (52) <0.001 
  3 49 (24) 19 (76) <0.001 30 (60) 19 (76) 0.170 
 Median operative time, min (IQR) 255 (221–300) 422 (278–532) <0.001 285 (253–335) 422 (278–531) 0.003 
 Blood products, n (%) 3 (1.4) 1 (4.3) 0.328 0 (0) 1 (4.3) 0.315 
 Anastomosis, n (%) 
  Manual 26 (11.5) 0 (0) 0.087 1 (2) 0 (0) >0.99 
  Stapler 201 (88.5) 25 (100) 0.087 49 (98) 25 (100) >0.99 
Postoperative 
 Chyle leak, n (%) 15 (7) 0 (0) 0.378 6 (12) 0 (0) 0.170 
 Complications, n (%) 65 (29) 4 (16) 0.019 17 (34) 4 (16) 0.009 
  Clavien-Dindo ≥3 26 (12) 3 (12) 0.371 6 (12) 3 (12) 0.406 
  Clavien-Dindo <3 54 (24) 6 (24) 0.090 16 (32) 6 (24) 0.704 
 Median LOS, days (IQR) 11 (8–19) 10 (7–12) 0.064 13.5 (8–25) 10 (7–12) 0.008 
 30-day mortality, n (%) 23 (10.1) 2 (8) >0.99 6 (12) 2 (8) 0.711 
 90-day mortality, n (%) 29 (12.8) 2 (8) 0.749 8 (16) 2 (8) 0.480 
Before matchingAfter matching
Variablesno patent (n = 227)patent (n = 25)p valueno patent (n = 50)patent (n = 25)p value
Surgical 
 Surgery type, n (%) 
  Trans-hiatal esophagectomy 45 (19.9) 0 (0) 0.011 2 (4) 0 (0) 0.550 
  Transthoracic esophagectomy 23 (10.2) 10 (40) <0.001 17 (34) 10 (40) 0.610 
  3-field esophagectomy 158 (69.9) 15 (60) 0.310 31 (62) 15 (60) >0.99 
 Surgery approach, n (%) 
  Laparoscopic/Robotic 179 (79.2) 22 (91.7) 0.182 48 (96) 23 (92) >0.99 
  Open 19 (43) 1 (4.2) 0.089 2 (4) 1 (4.2) >0.99 
 Surgeons, n (%) 
  1 174 (85.37) 10 (40) <0.001 34 (68) 10 (40) 0.020 
  2 136 (66.7) 13 (52) 0.147 45 (90) 13 (52) <0.001 
  3 49 (24) 19 (76) <0.001 30 (60) 19 (76) 0.170 
 Median operative time, min (IQR) 255 (221–300) 422 (278–532) <0.001 285 (253–335) 422 (278–531) 0.003 
 Blood products, n (%) 3 (1.4) 1 (4.3) 0.328 0 (0) 1 (4.3) 0.315 
 Anastomosis, n (%) 
  Manual 26 (11.5) 0 (0) 0.087 1 (2) 0 (0) >0.99 
  Stapler 201 (88.5) 25 (100) 0.087 49 (98) 25 (100) >0.99 
Postoperative 
 Chyle leak, n (%) 15 (7) 0 (0) 0.378 6 (12) 0 (0) 0.170 
 Complications, n (%) 65 (29) 4 (16) 0.019 17 (34) 4 (16) 0.009 
  Clavien-Dindo ≥3 26 (12) 3 (12) 0.371 6 (12) 3 (12) 0.406 
  Clavien-Dindo <3 54 (24) 6 (24) 0.090 16 (32) 6 (24) 0.704 
 Median LOS, days (IQR) 11 (8–19) 10 (7–12) 0.064 13.5 (8–25) 10 (7–12) 0.008 
 30-day mortality, n (%) 23 (10.1) 2 (8) >0.99 6 (12) 2 (8) 0.711 
 90-day mortality, n (%) 29 (12.8) 2 (8) 0.749 8 (16) 2 (8) 0.480 

LOS, length of stay, IQR, interquartile range.

No incidences of chyle leak were recorded in the study group, while 12% of patients in the control group experienced chyle leak (p = 0.17). Overall complication rate (16 vs. 34%, p = 0.009) and escalation of care (28 vs. 74%, p = 0.002) were significantly lower in the study group compared to the control. Patients who experienced postoperative chylothorax (15, none of which were in the study group) who did not self-resolve were initially attempted noninvasive treatment, including discontinuation of enteral feeding with/without use of a somatostatin analog, 3 patients required thoracentesis/additional thoracic drain insertion. Only 4 patients who experienced postoperative chyle leak were eventually reoperated. Two patients were successfully treated by thoracoscopic clipping of the thoracic duct; an additional patient was reoperated for attempted clipping, but due to failure to identify the thoracic duct and continuation of chylous leak, angiographic embolization was attempted and failed as well. Leak eventually resolved through prolonged fasting paired with use of somatostatin analog and total parenteral nutrition. An additional patient was reoperated not due to chyle leak, but for pancreatic leak resulting in hemorrhage and colonic necrosis, requiring two additional procedures, patient ultimately died of multiorgan failure. Postoperative length of stay was significantly shorter for the study group when compared to control (13.5 days [IQR 8–25] versus 10 days [IQR 7–12], p = 0.008).

Thoracic duct injury resulting in chylothorax is a major complication of esophagectomy that may require invasive procedures or re-intervention to treat it. This study demonstrated a 60% success rate for thoracic duct visualization through use of preoperative injection of patent blue. Few reports have addressed the issue of thoracic duct visualization to avoid injury to the duct (Table 4). Lubbers et al. [25] previously suggested intraoperative administration of a lipid-rich solution via feeding jejunostomy as a measure to visualize the thoracic duct intraoperatively in order to detect and treat inadvertent injury. This study compared a retrospective cohort of 59 patients who underwent Ivor-Lewis MIE and received lipid-rich solution intraoperatively to a historical cohort of identical patients who did not receive this solution. Thoracic duct was visible during dissection in 63% (37 patients) with intraoperative unintended duct damage detected and treated in 5% of patients (3/59 patients), compared to a 4.1% detection rate (3/74 patients) in the control group (p = 0.629). Postoperative chylothorax rate was 1.7% (1/59 patients) in the study. The current cohort of patients who received patent blue was recorded prospectively, offering greater validity to results when compared to a retrospective study. Moreover, while lipid-rich solution administration suggest a mode of detection for duct injury, patent blue visualization may offer prevention of such injury all together, further minimizing postoperative leak rates.

Table 4.

Previous studies for methods of thoracic duct visualization

AuthorYear publishedYears of studyPatients, nMethod% Visualized% Chyle leak
Shackcloth et al. [282001 1999–2000 93 Preoperative oral administration of single cream (3 h) for detection of the thoracic duct 100 
Shen et al. [272014 2006–2012 178 Preoperative oral administration of milk (6 h) for detection of the thoracic duct 95.5 0.56 
Oguma et al. [262018 2014–2017 130 Preoperative simulation of the thoracic duct using MRTD NA 1.5 
Du et al. [292018 2013–2016 192 Preoperative oral administration of olive oil (8 h) for detection of the thoracic duct 100 
Lubbers et al. [252019 2015–2016 59 Intraoperative administration of lipid-rich solution through feeding jejunostomy for detection of thoracic duct injury 5%  
Vecchiato et al. [242020 2018–2019 19 Intraoperative ICG injection percutaneously in the inguinal nodes for detection of the thoracic duct 100 
AuthorYear publishedYears of studyPatients, nMethod% Visualized% Chyle leak
Shackcloth et al. [282001 1999–2000 93 Preoperative oral administration of single cream (3 h) for detection of the thoracic duct 100 
Shen et al. [272014 2006–2012 178 Preoperative oral administration of milk (6 h) for detection of the thoracic duct 95.5 0.56 
Oguma et al. [262018 2014–2017 130 Preoperative simulation of the thoracic duct using MRTD NA 1.5 
Du et al. [292018 2013–2016 192 Preoperative oral administration of olive oil (8 h) for detection of the thoracic duct 100 
Lubbers et al. [252019 2015–2016 59 Intraoperative administration of lipid-rich solution through feeding jejunostomy for detection of thoracic duct injury 5%  
Vecchiato et al. [242020 2018–2019 19 Intraoperative ICG injection percutaneously in the inguinal nodes for detection of the thoracic duct 100 

ICG, indocyanine green; NA, not applicable; MRTD, magnetic resonance thoracic ductography.

One suggested method of preoperative duct imaging is simulation using magnetic resonance thoracic ductography (MRTD) [26]. A study by Oguma et al. found some type of abnormal finding on MRTD in 18.5% of patients (24/130). The rate of chylothorax was 5.6% (9/160 patients) in patients who underwent thoracoscopic esophagectomy without preoperative MRTD compared to 1.5% (2/130 patients) in patients who did; however, this difference was not significant (p = 0.133). In the current study, thoracic duct visualization was achieved in 60% of patients who underwent patent blue injection. No data were collected regarding duct abnormalities. Considering that variant anatomy in present 40–60% patients [38], it may be assumed that a certain percentage of patients for which the thoracic duct was visualized had normal duct anatomy. Cases in which duct identification was unsuccessful may be attributed to abnormal anatomy. Other possible explanations include post-chemoradiation tissue fibrosis, tumoral infiltration [39], or a less than optimal timing, volume, or concentration of the dye injection, as this is yet to be defined. As duct injury is not strictly reserved for patients with variant anatomy, intraoperative visualization using patent blue may offer greater anatomical identification and prevention of inadvertent injury when compared to preoperative use of MRTD for identification anatomical variants. The rate of postoperative chylothorax with use of patent blue in the current study was lower (0%) that with use of MRTD (2.5%). Differences were similarly nonsignificant when compared to control groups in both studies.

Other studies have suggested preoperative oral administration of cream [28], milk [27], or olive oil [29] to facilitate intraoperative duct injury identification, with promising results (range 95–100%). Other than varying prevalence of lactose intolerance in certain populations, the reason for differences in results may stem from the earlier administration of substrate in these studies (6–8 h prior to surgery as opposed to intraoperative administration).

Vechiatto et al. [24] explored use of indocyanine green (ICG) fluorescence as means for thoracic duct identification in 19 patients undergoing MIE in prone position. This study reported percutaneous bilateral injection of ICG in the superficial inguinal nodes using ultrasound visualization before thoracoscopy in total esophagectomy and after laparoscopic time in Ivor-Lewis esophagectomy. After which thoracoscopy was performed and near-infrared fluorescence was used to identify the thoracic duct and/or chyle leak. This study is limited by fact of need for additional imaging modalities in order to identify ICG and lack of a comparative group serving as a control. The thoracic duct was clearly identified in all 19 patients after a mean time of 52.7 min from time of injection, and no incidences of postoperative chyle leak were recorded. Optimal timing for duct identification using patent blue is yet to be defined and was not adequately explored in the current study.

To summarize, while cream [28], milk [27], or olive oil [29] may be utilized as a means for identification of intraoperative duct injury after the fact, this study attempts to identify the duct in order to avoid intraoperative injury. Suggested methods to improve duct visualization by use of patent blue may include modifications in injection technique, substance concentration, patients’ positioning, and timing from injection to thoracic stage of the procedure.

In our study, we did not find any difference in preoperative variables except for regimens used in neoadjuvant therapy, which we could not explain. Low preoperative albumin was previously suggested to be a risk factor postoperative chylothorax [8]. However, this factor did not differ between groups in this study’s cohort. Ohkora et al. [3] suggested that post-chemoradiotherapy (HR = 3.43) is an independent factor predicting chylothorax. The current cohort did demonstrate significant differences in rate of receipt of CROSS protocol (33.3 vs. 6%, p = 0.004). However, overall rates of receipt of preoperative radiotherapy did not differ between groups (56 vs. 60%, p = 0.087).

Previous data suggest that patients with a BMI <25 have a higher risk of developing chylothorax, suggesting as explanation that as BMI increases, more fatty tissue surrounds the tumor and esophagus, leading to better protection of the thoracic duct [4, 7]. Median BMI was 25.9 kg/m2 (IQR 22.1–25.9) in the current study group and 23.4 kg/m2 (IQR 20.8–26.7) in the control group, with no significant differences (p = 0.32). This suggests the possible utility for of patent blue for visualization in patients with a lower BMI. Nonetheless, current small cohort size limits the ability to demonstrate any significant differences in chylothorax rates.

Squamous tumor type has been previously suggested to be an independent prognostic factor for post-esophagectomy chylothorax [5, 6]. The current study used squamous pathology for PS matching, limiting any confounding effect it may have had on rates of postoperative chylothorax.

Tumor location and extent and location of lymph node metastasis were previously identified as risk factors for post-esophagectomy chylothorax as well [40]. This is thought to result from difficult mediastinal dissection in middle thoracic esophageal cancer leading to a higher thoracic duct injury risk. Tumor locations did not significantly differ between study and control groups after PS matching. Nguyen et al. [8] previously suggested number of harvested lymph nodes as a significant predictor of chylothorax following esophagectomy. Median number of harvested lymph nodes in the study group was 19 (IQR 13–29.5), and while it was significantly higher compared to other esophagectomies before PS matching (13 nodes [IQR 9–18], p < 0.001), this did not remain significant after matching (18 nodes [12–23], p = 0.366).

Longer median operative times for the study group both before (255 vs. 422 min, p < 0.001) and after matching (422 vs. 285 min, p = 0.003) may be explained by procedure prolongation due not only to only patent blue injection but perhaps additional efforts/time spent by operating surgeons in attempts to locate the dye in the mediastinum.

This study holds several limitations, including small cohort size limiting ability to achieve significant results. Furthermore, data pertaining to the control group were collected retrospectively, further limiting comparisons.

In conclusion, our study suggests a new and simple method of thoracic duct visualization during MIE in an attempt to decrease chyle leak rates and enhance good surgical outcomes. This method may offer an easy, affordable, and reproducible solution for duct visualization in the attempt to avoid duct injury during esophagectomy. Further studies, preferably in a randomized controlled setting, are required in order to better characterize the utility and feasibility of this method in preventing chylothorax.

The authors wish to thank Nisim Mery for his assistance and guidance with statistical analysis.

The research has been carried out within an appropriate ethical framework. Written informed consent was obtained for participation. Protocol and consent forms were approved by the Institutional Review Board at Rabin Medical Center (0489-17-RMC).

The authors of this manuscript have no conflicts of interest to disclose.

No external funding was received for this paper.

Yael Berger (corresponding author): design of work, data acquisition, analysis and interpretation, and manuscript drafting. Vyacheslav Bard (#equal contribution as 1st author): conception and design of the work, data acquisition, result interpretation, manuscript drafting, and critical review for important intellectual content. Muhammad Abbas and Daniel Solomon: substantial contribution in data acquisition. Nikolai Menasherov: substantial contributions to data acquisition and reviewing it critically for important intellectual content. Hanoch Kashtan: substantial contributions to data acquisition and result interpretation, reviewing it critically for important intellectual content, and final approval of the version to be published.

Additional Information

Yael Berger and Vyacheslav Bard contributed equally to this work and are co-first authors.

The datasets generated and/or analyzed during the current study are not publicly available due to patient confidentiality and institutional regulations but are readily available from the corresponding author on reasonable request.

1.
Mariette
C
,
Piessen
G
,
Triboulet
JP
.
Therapeutic strategies in oesophageal carcinoma: role of surgery and other modalities
.
Lancet Oncol
.
2007
;
8
(
6
):
545
53
.
2.
Tverskov
V
,
Wiesel
O
,
Solomon
D
,
Orgad
R
,
Kashtan
H
.
The impact of cervical anastomotic leak after esophagectomy on long-term survival of patients with esophageal cancer
.
Surgery
.
2021
[cited 2021 Nov 22]; Available from: https://pubmed.ncbi.nlm.nih.gov/34750016.
3.
Ohkura
Y
,
Ueno
M
,
Shindoh
J
,
Iizuka
T
,
Ka
H
,
Udagawa
H
.
Risk factors for postoperative chylothorax after radical subtotal esophagectomy
.
Ann Surg Oncol
.
2018
;
25
(
9
):
2739
46
.
4.
Miao
L
,
Zhang
Y
,
Hu
H
,
Ma
L
,
Shun
Y
,
Xiang
J
, et al
.
Incidence and management of chylothorax after esophagectomy
.
Thorac Cancer
.
2015
;
6
(
3
):
354
8
.
5.
Shah
RD
,
Luketich
JD
,
Schuchert
MJ
,
Christie
NA
,
Pennathur
A
,
Landreneau
RJ
, et al
.
Postesophagectomy chylothorax: incidence, risk factors, and outcomes
.
Ann Thorac Surg
.
2012
;
93
(
3
):
897
904
.
6.
Kranzfelder
M
,
Gertler
R
,
Hapfelmeier
A
,
Friess
H
,
Feith
M
.
Chylothorax after esophagectomy for cancer: impact of the surgical approach and neoadjuvant treatment – systematic review and institutional analysis
.
Surg Endosc
.
2013
;
27
(
10
):
3530
8
.
7.
Zhang
SS
,
Yang
H
,
Luo
KJ
,
Huang
QY
,
Chen
JY
,
Yang
F
, et al
.
The impact of body mass index on complication and survival in resected oesophageal cancer: a clinical-based cohort and meta-analysis
.
Br J Cancer
.
2013
;
109
(
11
):
2894
903
.
8.
Nguyen
D
,
Finch
R
,
O’Loughlin
B
,
Dodd
BR
.
Incidence, risk factors and management of chylothorax post-oesophagectomy for oesophageal carcinoma in an Australian institution
.
Ann Esophagus
.
2020
;
3
(
0
):
3
.
9.
Yoshida
N
,
Baba
H
.
Risk factors for chylothorax after esophagectomy
.
J Thorac Dis
.
2019
;
11
(
Suppl 3
):
S196
7
.
10.
Malibary
N
,
Manfredelli
S
,
Almuttawa
A
,
Delhorme
J-B
,
Romain
B
,
Brigand
C
, et al
.
Evaluating the surgeon’s experience as a risk factor for post-esophagectomy chylothorax on a four-year cohort
.
Cureus
.
2020
;
12
(
6
):
e8696
.
11.
Marthaller
KJ
,
Johnson
SP
,
Pride
RM
,
Ratzer
ER
,
Hollis
HW
.
Percutaneous embolization of thoracic duct injury post-esophagectomy should be considered initial treatment for chylothorax before proceeding with open re-exploration
.
Am J Surg
.
2015
;
209
(
2
):
235
9
.
12.
Carcoforo
P
,
Soliani
G
,
Maestroni
U
,
Donini
A
,
Inderbitzin
D
,
Hui
TT
, et al
.
Octreotide in the treatment of lymphorrhea after axillary node dissection: a prospective randomized controlled trial
.
J Am Coll Surg
.
2003
;
196
(
3
):
365
9
.
13.
Tachibana
M
,
Kinugasa
S
,
Yoshimura
H
,
Dhar
DK
,
Ueda
S
,
Fujii
T
, et al
.
Does fibrin glue reduce lymph leakage (pleural effusion) after extended esophagectomy? Prospective randomized clinical trial
.
World J Surg
.
2003
;
27
(
7
):
776
81
.
14.
Ohkura
Y
,
Ueno
M
,
Iizuka
T
,
Haruta
S
,
Tanaka
T
,
Udagawa
H
.
New combined medical treatment with etilefrine and octreotide for chylothorax after esophagectomy a case report and review of the literature
.
Med
.
2015
;
94
(
49
):
e2214
.
15.
Tamura
T
,
Kubo
N
,
Yamamoto
A
,
Sakurai
K
,
Toyokawa
T
,
Tanaka
H
, et al
.
Cervical chylous leakage following esophagectomy that was successfully treated by intranodal lipiodol lymphangiography: a case report
.
BMC Surg
.
2017
;
17
(
1
):
20
.
16.
Lin
Y
,
Li
Z
,
Li
G
,
Zhang
X
,
Deng
H
,
Yang
X
, et al
.
Selective en masse ligation of the thoracic duct to prevent chyle leak after esophagectomy
.
Ann Thorac Surg
.
2017
;
103
(
6
):
1802
7
.
17.
Ohkura
Y
,
Ueno
M
,
Iizuka
T
,
Udagawa
H
.
Effectiveness of etilefrine regimen for chylothorax after esophagectomy with thoracic duct resection
.
Esophagus
.
2018
;
15
(
1
):
33
8
.
18.
Gupta
R
,
Singh
H
,
Kalia
S
,
Gupta
R
,
Singh
R
,
Verma
GR
.
Chylothorax after esophagectomy for esophageal cancer: risk factors and management
.
Indian J Gastroenterol
.
2015
;
34
(
3
):
240
4
.
19.
Alamdari
DH
,
Asadi
M
,
Rahim
AN
,
Maddah
G
,
Azizi
S
,
Shahidsales
S
, et al
.
Efficacy and safety of pleurodesis using platelet-rich plasma and fibrin glue in management of postoperative chylothorax after esophagectomy
.
World J Surg
.
2018
;
42
(
4
):
1046
55
.
20.
Guo
W
,
Zhao
YP
,
Jiang
YG
,
Niu
HJ
,
Liu
XH
,
Ma
Z
, et al
.
Prevention of postoperative chylothorax with thoracic duct ligation during video-assisted thoracoscopic esophagectomy for cancer
.
Surg Endosc
.
2012
;
26
(
5
):
1332
6
.
21.
Lei
Y
,
Feng
Y
,
Zeng
B
,
Zhang
X
,
Chen
J
,
Zou
J
, et al
.
Effect of prophylactic thoracic duct ligation in reducing the incidence of postoperative chylothorax during esophagectomy: a systematic review and meta-analysis
.
Thorac Cardiovasc Surg
.
2018
;
66
(
5
):
370
5
.
22.
Schurink
B
,
Mazza
E
,
Ruurda
JP
,
Roeling
TAP
,
Steenhagen
E
,
Bleys
RLAW
, et al
.
Low-fat tube feeding after esophagectomy is associated with a lower incidence of chylothorax
.
Ann Thorac Surg
.
2019
;
108
(
1
):
184
9
.
23.
Glatz
T
,
Marjanovic
G
,
Hoeppner
J
.
[Prevention and surgical therapy of chylothorax]
.
Zentralbl Chir
.
2018
;143(
03
):
278
83
.
24.
Vecchiato
M
,
Martino
A
,
Sponza
M
,
Uzzau
A
,
Ziccarelli
A
,
Marchesi
F
, et al
.
Thoracic duct identification with indocyanine green fluorescence during minimally invasive esophagectomy with patient in prone position
.
Dis Esophagus
.
2020
;
33
(
12
):
doaa030
6
.
25.
Lubbers
M
,
van Det
M
,
Kouwenhoven
E
.
Intraoperative lipid-rich nutrition in the detection of chylothorax in minimally invasive ivor lewis esophagectomy
.
Surg Innov
.
2019
;
26
(
5
):
545
50
.
26.
Oguma
J
,
Ozawa
S
,
Kazuno
A
,
Nitta
M
,
Ninomiya
Y
,
Yatabe
K
, et al
.
Clinical significance of new magnetic resonance thoracic ductography before thoracoscopic esophagectomy for esophageal cancer
.
World J Surg
.
2018
;
42
(
6
):
1779
86
.
27.
Shen
Y
,
Feng
M
,
Khan
MA
,
Wang
H
,
Tan
L
,
Wang
Q
.
A simple method minimizes chylothorax after minimally invasive esophagectomy
.
J Am Coll Surg
.
2014
;
218
(
1
):
108
12
.
28.
Shackcloth
MJ
,
Poullis
M
,
Lu
J
,
Page
RD
.
Preventing of chylothorax after oesophagectomy by routine pre-operative administration of oral cream
.
Eur J Cardio Thorac Surg
.
2001
;
20
(
5
):
1035
6
.
29.
Du
ZS
,
Li
XY
,
Luo
HS
,
Wu
SX
,
Zheng
CP
,
Li
ZY
, et al
.
Preoperative administration of olive oil reduces chylothorax after minimally invasive esophagectomy
.
Ann Thorac Surg
.
2019
;
107
(
5
):
1540
3
.
30.
Dindo
D
,
Demartines
N
,
Clavien
P
.
Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey
.
Ann Surg
.
2004
;
240
(
2
):
205
13
.
31.
Olivier
F
,
Courtois
A
,
Jossa
V
,
Bruck
G
,
Aouachria
S
,
Coibion
M
, et al
.
Sentinel lymph node mapping with patent blue dye in patients with breast cancer: a retrospective single institution study
.
Gland Surg
.
2021
;
10
(
9
):
2600
7
.
32.
Korbi
A
,
Khaskhoussy
A
,
Cherif
O
,
Hajji
A
,
Gaddab
I
,
Chaabene
A
, et al
.
Anaphylactic shock due to patent blue: case report and review of literature
.
Pan Afr Med J
.
2018
;
31
:
7
.
33.
Nusrath
S
,
Thammineedi
SR
,
Saksena
AR
,
Patnaik
SC
,
Reddy
P
,
Usofi
Z
, et al
.
Thoracic duct lymphography by near-infrared indocyanine green fluorescence imaging in thoracic surgery. A review
.
Indian J Surg Oncol
.
2022
;
13
(
2
):
415
20
.
34.
Haukoos
JS
,
Lewis
RJ
.
The propensity score
.
JAMA
.
2015
;
314
(
15
):
1637
8
.
35.
Siewert
JR
,
Stein
HJ
.
Carcinoma of the gastroesophageal junction: classification, pathology and extent of resection
.
Dis Esophagus
.
1996
;
9
(
3
):
173
82
. Available from: https://academic.oup.com/dote/article/9/3/173/2798630.
36.
Barthelmes
L
,
Goyal
A
,
Newcombe
RG
,
McNeill
F
,
Mansel
RE
;
NEW START and ALMANAC Study Groups
.
Adverse reactions to patent blue V dye: the NEW START and ALMANAC experience
.
Eur J Surg Oncol
.
2010
;
36
(
4
):
399
403
.
37.
Suzuki
Y
,
Yamana
K
,
Ishizawa
H
,
Kawai
H
,
Matsuda
Y
,
Kato
R
, et al
.
Successful thoracic duct identification with patent blue V during thoracic duct ligation for chylothorax: a case report
.
Cases Gen Thorac Cardiovasc Surg
.
2022
;
1
(
1
):
7
.
38.
Toliyat
M
,
Singh
K
,
Sibley
RC
,
Chamarthy
M
,
Kalva
SP
,
Pillai
AK
.
Interventional radiology in the management of thoracic duct injuries: anatomy, techniques and results
.
Clin Imaging
.
2017
;
42
:
183
92
.
39.
Varshney
VK
,
Nayar
R
,
Soni
SC
,
Selvakumar
B
,
Garg
PK
,
Varshney
P
, et al
.
Intra-nodal indocyanine green injection to delineate thoracic duct during minimally invasive esophagectomy
.
J Gastrointest Surg
.
2022
;
26
(
8
):
1559
65
.
40.
Fujita
T
,
Sato
T
,
Sato
K
,
Hirano
Y
,
Fujiwara
H
,
Daiko
H
.
Clinical manifestation, risk factors and managements for postoperative chylothorax after thoracic esophagectomy
.
J Thorac Dis
.
2019
;
11
:
S198
201
.