Patient Population and Circulating Cell Presence:
In total, 32 stage I-III newly diagnosed untreated locally advanced EC patients were recruited prior to receiving standard of care CRT, with CRT regimes averaging 37 days of radiation. A minimum of one CAML was found in 88% (n = 28/32) of patient samples, including all available time points throughout treatment. In line with standard EC demographics, this study’s cohort consisted of 91% (n = 29/32) male and 9% (n = 3/32) female patients (Table 1). EAC accounted for 78% of patients (n = 25/32) and SCC 22% (n = 7/32). Clinical stage distribution was found as such, 3% (n = 1/32) in stage Ib, 3% (n = 1/32) in stage Ic, 12% (4/32) in stage IIa, 19% (n = 6/32) in stage IIb, 44% in stage IIIa, 16% (n = 5/32) in stage IIIb, and 3% (n = 1/32) of unknown stage with no evidence of metastasis.
In order to evaluate the relationship between CAML size and patient outcomes, PFS and OS based on CAML sizes ≥ 50 µm or < 50 µm were compared at BL, T1 and T2. The enumeration of CAML cells found an average of 6.5 CAMLs/7.5 mL of blood throughout all known time points, with an average CAML size of 47 µm. Including all 77 available time points used in this study, CAML monitoring via microfiltration had a fail rate of 3% (n = 2/77) due to clotting of the blood. CAML identification was based on a large cell diameter, polyploid nuclei, and high expression of CD45. Average CAML distribution by clinical stage can be found in Supplementary Table 1.
In addition to CAMLs, CTC presence was found in only 12.5% of the samples (n = 4/32) throughout all available time points, with two patients presenting stage IIa, one patient stage IIb, and the fourth patient of unknown stage. CTCs were identified by their mononucleated appearance, high expression of cytokeratin, and lack of CD45 expression. To compare the differences between circulating cancer-associated cells, Fig. 1 portrays a size and stain comparison between a CAML, CTC, and a regular white blood cell.
Sequential CAML Monitoring:
At BL, prior to the start of CRT (n = 29), we found an average of 5 CAMLs/7.5 mL of blood, with a minimum of one CAML seen in 76% (n = 22/29) of available baseline samples. The average max CAML size seen among patients at BL was found to be 38 µm. When comparing survival outcomes based on CAML size, patients with CAML sizes < 50 µm (n = 23) had non-significant trends toward improved PFS compared to patients with CAML sizes ≥ 50 µm (n = 6) (HR = 5.3, 95%CI = 1.0-27.7, p = 0.190). (Fig. 2a) Similarly, OS at BL showed that patients with CAML sizes < 50 µm trended toward improved survival over patients with CAMLs ≥ 50 µm (HR = 8.5, 95%CI = 1.4–51.3, p = 0.060) (Fig. 2b).
At T1, the midpoint of radiation therapy, we found an average of 10 CAMLs/7.5 mL of blood, with an average CAML size of 57 µm, and CAMLs were found in 96% (n = 23/24) of samples. Despite most samples having CAMLs at this time point, CAML size was not a significant prognostic indicator when comparing the < 50 µm group (n = 12) versus the ≥ 50 µm group (n = 12); PFS (HR = 1.2, 95%CI = 0.371–4.068, p = 0.974) and OS (HR = 1.6, 95%CI = 0.5–5.6, p = 0.678). (Figs. 2c,d) At this time, these initial findings indicate that there appears to be no statistical clinical significance of CAML size and presence at T1.
At T2, the first sample taken after the completion of radiation therapy, we found an average of 5 CAMLs/7.5 mL of blood and an average diameter of 46 µm was identified. In all available samples at this time point, CAMLs were found in 89% (n = 17/19) of patients. We found that patients with < 50 µm CAMLs had significantly improved PFS and OS when compared patients with ≥ 50 µm CAMLs, PFS (HR = 12.0, 95%CI = 2.7–54.1, p = 0.004) and OS (HR = 9.0, 95%CI = 1.9–43.5, p = 0.019). (Figs. 2e,f) However, given the small patient population in this analysis (n = 19) expanded patient population is necessary.
Patterns Found In Locally Advanced EC:
To better understand the stratification of survival based on CAML size, we evaluated the two primary patient populations- patients that eventually progressed within 24 months and patients that did not progress within 24 months (Fig. 3). After completion of radiation (T2), 85% (n = 11/13) of non-progressing patients had smaller CAMLs (< 50 µm). In contrast, only 10% (n = 1/10) of patients that progressed had smaller CAMLs (< 50 µm). Using 24 months as an endpoint, this would equate to a prognostic accuracy of 87% in predicting disease recurrence based on the presence of large CAMLs after completion of radiation.
Further evaluation of CAML trends during treatment seemed to identify two general patterns in CAMLs engorgement. In patients that eventually progressed, average CAML sizes increased from 51.7 µm (BL), to 58.7 µm (T1), and 67.1 µm (T2). This demonstrated a linear growth in CAML engorgement from baseline to the completion of CRT. In patients that did not progress average CAML sizes increased from 37.4 µm (BL) to 58.5 µm (T1). However, by T2 average CAMLs decreased to 41.8 µm. In total while most patients (71%) had an increase from BL to T1, patients, however in the patients that progressed there was an additional 14% increase in CAML size at T2. In contrast, in patients that did not progress there was a 29% decrease at the T2 blood sampling. While further analyses in a larger patient population will be needed to confirm these trends in CAML growth in relation to PFS and OS, this pattern does suggest a biological and clinical difference in a patient’s immunological response to radiation treatment.
In addition to CAML analysis, patients’ OS and PFS were analyzed based on EC histology and whether or not they had received surgical resection (Table 1). When comparing the two histological types, we found that patients with EAC (n = 25) trended for improved PFS over SCC (n = 7) (HR = 3.9, 95%CI = 0.91–14.8, p = 0.149). Interestingly, OS was significantly in favor for EAC (HR = 14.8, 95%CI = 2.63–83.33, p = 0.009). In line with previous studies, patients that had received surgery post-completion of standard CRT trended toward improved PFS compared to patients that did not undergo surgery, PFS (HR = 2.8, 95%CI = 0.12–1.04, p = 0.104) but OS did not appear different (HR = 2.5, 95%CI = 0.12–1.32, p = 0.228).