Endoscopic therapy is currently the standard of care for the treatment of high-grade dysplasia (HGD) or intramucosal adenocarcinoma (IMC) in patients with Barrett's esophagus (BE). Visible lesions are treated with endoscopic mucosal resection (EMR), which is often coupled with radiofrequency ablation (RFA). However, endoscopic therapy may require multiple sessions (one session every 2-3 months) and does not always assure complete eradication of neoplasia. Furthermore, despite complete eradication, recurrences are not uncommon. This study assesses which potential risk factors can predict a poor response after endoscopic sessions. Forty-five BE patients who underwent at least one endoscopic session (EMR alone or ablation with or without preceding EMR) for the treatment of HGD/IMC, low-grade dysplasia (LGD), or indefinite for dysplasia (IND) were analyzed. DNA flow cytometry was performed on 82 formalin-fixed paraffin-embedded samples from the 45 patients, including 78 HGD/IMC, 2 LGD, and 2 IND. Eight non-dysplastic BE samples were used as controls. Three to four 60-micron thick sections were cut from each tissue block, and the area of HGD/IMC, LGD, or IND was manually dissected. Potential associations between clinicopathologic risk factors and persistent/recurrent HGD/IMC following each endoscopic session were examined using univariate and multivariate Cox models with frailty terms. Sixty (73%) of the 82 specimens showed abnormal DNA content (aneuploidy or elevated 4N fraction). These were all specimens with HGD/IMC (representing 77% of that group). Of these 60 HGD/IMC samples with abnormal DNA content, 42 (70%) were associated with subsequent development of persistent/recurrent HGD/IMC (n = 41) or esophageal adenocarcinoma (EAC; n = 1) within a mean follow-up time of 16 months (range: 1 month to 9.4 years). In contrast, only 6 (27%, all HGD/IMC) of the 22 remaining samples (all with normal DNA content) were associated with persistent/recurrent HGD/IMC. For outcome analysis per patient, 11 (24%) of the 45 patients developed persistent/recurrent HGD/IMC or EAC, despite multiple endoscopic sessions (mean: 3.6, range: 1-11). In a univariate Cox model, the presence of abnormal DNA content (hazard ratio [HR] = 3.8, p = 0.007), long BE segment ≥ 3 cm (HR = 3.4, p = 0.002), endoscopic nodularity (HR = 2.5, p = 0.042), and treatment with EMR alone (HR = 2.9, p = 0.006) were significantly associated with an increased risk for persistent/recurrent HGD/IMC or EAC. However, only abnormal DNA content (HR = 6.0, p = 0.003) and treatment with EMR alone (HR = 2.7, p = 0.047) remained as significant risk factors in a multivariate analysis. Age ≥ 60 years, gender, ethnicity, body mass index (BMI) ≥ 30 kg/m, presence of hiatal hernia, and positive EMR lateral margin for neoplasia were not significant risk factors for persistent/recurrent HGD/IMC or EAC (p > 0.05). Three-month, 6-month, 1-year, 3-year, and 6-year adjusted probabilities of persistent/recurrent HGD/IMC or EAC in the setting of abnormal DNA content were 31%, 56%, 67%, 79%, and 83%, respectively. The corresponding probabilities in the setting of normal DNA content were 10%, 21%, 28%, 38%, and 43%, respectively. In conclusion, in BE patients with baseline HGD/IMC, both DNA content abnormality and treatment with EMR alone were significantly associated with persistent/recurrent HGD/IMC or EAC following each endoscopic session. DNA content abnormality as detected by DNA flow cytometry identifies HGD/IMC patients at highest risk for persistent/recurrent HGD/IMC or EAC, and it also serves as a diagnostic marker of HGD/IMC with an estimated sensitivity of 77%. The diagnosis of HGD/IMC in the setting of abnormal DNA content may warrant alternative treatment strategies as well as long-term follow-up with shorter surveillance intervals.