Lung cancer screening programs can contribute significantly to the incidental identification of esophageal cancer in a high-risk population

Introduction

Esophageal cancer, comprising about 1% of all cancer cases in the United States, is relatively rare, with an incidence of 4/100,000. ¹ It is most commonly diagnosed at later stages when lesions have enlarged to the point where the patient becomes symptomatic with dysphagia or unexplained weight loss. Smoking is a major risk factor for developing esophageal cancer.

Lung cancer screening (LCS) is broadly accepted and utilized in patients with at least a 20 pack/year smoking history between the ages of 50 and 80. LCS often identifies incidental findings (S-findings) such as cardiac findings or other neoplasms in the thorax. We report the magnitude of incidentally found esophageal cancers during our initial 4 years of LCS in a VAMC (Veteran's Administration Medical Center) population.

Material and methods

As a quality initiative exempt from IRB (Institutional Review Board) oversight, a retrospective database and chart review of a single-center VAMC LCS program in the underserved Western North Carolina region of Appalachia was performed. Esophageal S-findings (distal esophageal thickening, dilated esophagus, retained material and paraoesophageal lymph nodes) of asymptomatic patients with subsequently diagnosed esophageal adenocarcinoma cancers were recorded and compared to the symptomatically detected adenocarcinomas within the healthcare system.

Results

During the initial 4 years of LCS, 3893 patients were screened in the program. Ten asymptomatic esophageal cancers were found by LCS with a prevalence of 0.26%. These asymptomatic esophageal cancers accounted for 19% of our entire healthcare system's detected adeno-esophageal cancers (53/48,174). During the same time period, the prevalence of symptomatically detected cancers was 0.097% (43/44,452; Table 1). The prevalence of the LCS-detected asymptomatic cancers was 2.7× higher than the prevalence of symptomatic cancers (p = 0.004). Comparing the prevalence of the VAMC's symptomatically detected and LCS-detected esophageal cancers to the incidence of esophageal cancers in the general population during this same 4-year time frame (0.016%), ¹ our symptomatically detected cancer prevalence was significantly higher by a factor of 6× (p = 0.0001). Our LCS-detected asymptomatic cancer prevalence was even higher at 16× (p = 0.0001); however, we are limited to comparing our asymptomatic cancer prevalence to the incidence of only symptomatic cancers in the general population, since there is no established screening program for asymptomatic esophageal cancers and thus no data available.

OPEN IN VIEWER

Table 1.

Patient and case characteristics.

Characteristic	Value
Total patients, N (%)	53
Asymptomatic LCS group	10 (19%)
Symptomatic group	43 (81%)
Age in years, median (range)	72.0 (59–90)
Asymptomatic LCS group	71.0 (66–80)
Symptomatic group	72.2 (59–90)
Sex, N (%)
Male	52 (98%)
Female	1 (2%) (symptomatic group)
Race, N (%)
White	48 (90%)
African American	1 (2%)
Prefer not to share	4 (8%)
Esophageal cancer stage at time of diagnosis, N (%)
Asymptomatic LCS group (9)
I	1 (10%)
IIA	2 (20%)
IIB	1 (10%)
III	5 (50%)
IV	1 (10%)
Symptomatic group (43)
I	0
IIA	4 (9%)
IIB	3 (7%)
III	17 (40%)
IV	18 (42%)
Unstageable	1 (2%)
Smoking pack/year history in patients with esophageal adenocarcinoma:	Mean, years (standard deviation)
Asymptomatic LCS group (10)	46.4 (17.6)
Symptomatic + Asymptomatic LCS group (14)	51.7 (20.1)
Non-LCS group (39)	22.9 (21.8)
Total patients undergoing EGD during study time	665
LCS prompted EGDs due to LDCT findings	69 (10.4% of total)
Asymptomatic LCS detected esophageal cancers	10 (14.5% of LCS prompted)
Benign	38
Reflux disease	11
Barrett's disease	6
Stricture	3
Other	11

LCS: lung cancer screening; LDCT: low-dose computed tomography: EGD: esophagogastroduodenoscopy. Data are reported as n (%) unless otherwise noted.

The LCS group showed a significantly higher pack/year smoking history (52 vs 27 pack/years, p = 0.004), likely a factor for the higher esophageal cancer prevalence in our Veteran population, which was more than 6.9× the prevalence found in the average adult population. The incidentally LCS-detected asymptomatic esophageal cancers showed a significantly lower stage at time of detection (p = 0.025, Figure 1). This was most impressive in the number of stage IV cases found: 10% (1/10) in the screening-detected group vs. 43% (18/42) in the symptomatically-detected group. This showed a trend towards significance of p = 0.052.

OPEN IN VIEWER

Figure 1.

Esophageal cancer case frequency and stage by detection type. Esophageal cancer found in asymptomatic patients undergoing LDCT was diagnosed at earlier stages on average than in patients diagnosed during work-up for dysphagia.

New esophageal S-findings on low-dose computed tomography (LDCT) prompted advice to undergo an esophagogastroduodenoscopy (EGD). If patients had chronic esophageal disease (e.g. Barrett's or gastroesophageal reflux disease, GERD) and were on continuous surveillance, no new EGD was ordered. Of 72 patients with new and concerning esophageal findings, 69 agreed after gastroenterology consultation to undergo an EGD. This high adherence rate is likely due to our integrated VA Healthcare system and trusted personal communication by our LDCT nurse-navigator team. The three patients who refused workup will continue to be evaluated within the LCS program. These 69 EGDs accounted for 10.4% of the 665 Veterans undergoing EGDs during the 4-year period in our healthcare system; 14.5% (or 1 in 7) of the LDCT-initiated EGDs detected an asymptomatic esophageal cancer.

Post-hoc CT-scan analysis revealed that esophageal wall thickness of 7 mm or greater, paraoesophageal or gastric/paracardial lymph nodes greater than or equal to 7.0 mm, and retained esophageal content are strongly correlated to finding esophageal cancer during EGD (p < 0.001). The combination of these three CT-scan findings shows a sensitivity of 65% and specificity 100%, with a positive predictive value of 100% and a negative predictive value of 86%.

Discussion

Patients in LCS programs are at greatly increased risk of esophageal cancer, and higher vigilance for esophageal findings would likely result in earlier diagnosis and treatment. After our facility noticed this trend, our radiologists began to take special care upon reading LDCT scans to identify esophageal thickening and para-esophageal lymph nodes (Supplemental Figure). We communicated with primary care providers to discuss low thresholds for EGD for esophageal thickening or dysphagia symptoms in this population. Our LCS Nurse Practitioner calls these patients to explain the EGD procedure and stresses the importance of completing their testing. Over a 3.5-year time frame, we found six asymptomatic cases, but after increased vigilance, beginning mid-2024, the four additional cases were found in just 6 months.

It is possible the higher rates of esophageal cancer are at least partially unique to our Veteran population, being mostly males (a known 4× higher risk than in females ² ), and with many having known exposure risk factors in service, such as combustion byproducts and asbestos. ³ Agent orange and burn pits are also known exposures in our cohort and though not proven risk factors, this could be an area for future research.

The natural history of esophageal cancers is not well understood, and it is not clear how many early-stage esophageal cancers will go on to become symptomatic, especially since there is no dedicated screening established. Incidental esophageal findings during LCS could become a steppingstone in early detection.

In the Early Lung Cancer Action Project (ELCAP) study, only two esophageal cancers were detected out of 9263 baseline studies. ⁴ It is possible these figures would be higher with focused imaging analysis of esophageal findings. ⁵ The literature is scarce on incidental esophageal findings during LCS, consecutive work-up, and diagnosis of esophageal cancer; it is often only available as an abstract submitted to or presented at a meeting.^6,7

Conclusion

LCS in a Veteran population identified esophageal cancers 2.7× more often than in unscreened Veterans. In Veterans qualifying for LCS, the prevalence of esophageal cancer was 6.9× higher compared to the general population. About 1/389 Veterans undergoing LCS was found to have an asymptomatic esophageal cancer. LCS detected esophageal cancer at a lower stage, and the patients presented less often at stage 4. LCS programs should be vigilant regarding incidental esophageal findings, which can result in an earlier stage detection of asymptomatic esophageal cancer patients.

Supplemental Material