A nomogram associated with high probability of malignant nodes in the surgical specimen after trimodality therapy of patients with oesophageal cancer

Abstract

Background: The presence of malignant lymph nodes (+ypNodes) in the surgical specimen after preoperative chemoradiation (trimodality) in patients with oesophageal cancer (EC) portends a poor prognosis for overall survival (OS) and disease-free survival (DFS). Currently, none of the clinical variables highly correlates with +ypNodes. We hypothesised that a combination of clinical variables could generate a model that associates with high likelihood of +ypNodes after trimodality in EC patients.

Methods: We report on 293 consecutive EC patients who received trimodality therapy. A multivariate logistic regression analysis that included pretreatment and post-chemoradiation variables identified independent variables that were used to construct a nomogram for +ypNodes after trimodality in EC patients.

Results: Of 293 patients, 91 (31.1%) had +ypNodes. OS (p=0.0002) and DFS (p<0.0001) were shorter in patients with +ypNodes compared to those with -ypNodes. In multivariable analysis, the significant variables for +ypNodes were: baseline T-stage (odds ratio [OR], 7.145; 95% confidence interval [CI], 1.381-36.969; p=0.019), baseline N-stage (OR, 2.246; 95% CI, 1.024-4.926; p=0.044), tumour length (OR, 1.178; 95% CI, 1.024-1.357; p=0.022), induction chemotherapy (OR, 0.471; 95% CI, 0.242-0.915; p=0.026), nodal uptake on post-chemoradiation positron emission tomography (OR, 2.923; 95% CI, 1.007-8.485; p=0.049) and enlarged node(s) on post-chemoradiation computerised tomography (OR, 3.465; 95% CI, 1.549-7.753; p=0.002). The nomogram after internal validation using the bootstrap method (200 runs) yielded a high concordance index of 0.756.

Conclusion: Our nomogram highly correlates with the presence of +ypNodes after chemoradiation, however, considerably more refinement is needed before it can be implemented in the clinic.

Figure1

Sensitivity and Specificity of post chemoradiation PET and CT for lymph node metastasis. Sensitivities of both imaging are low (21.6% and 41.6%, respectively), whereas specificities are high (93.7% and 85.7%, respectively).

Figure1

Sensitivity and Specificity of post chemoradiation PET and CT for lymph node metastasis. Sensitivities of both imaging are low (21.6% and 41.6%, respectively), whereas specificities are high (93.7% and 85.7%, respectively).

Figure2

Nomogram to predict the probability of pathological lymph node metastasis for esophageal cancer patients using pretreatment and post chemoradiation clinical values. (Top) Points are obtained according to prognostic contribution of parameters. (Bottom) Points are translated to probability of pathological lymph node metastasis. The total points score is obtained by summing the points for each covariate of the nomogram which are baseline T, baseline N, tumor length, induction chemo, lymph node metastasis by post chomoradiation PET and CT.

Figure2

Nomogram to predict the probability of pathological lymph node metastasis for esophageal cancer patients using pretreatment and post chemoradiation clinical values. (Top) Points are obtained according to prognostic contribution of parameters. (Bottom) Points are translated to probability of pathological lymph node metastasis. The total points score is obtained by summing the points for each covariate of the nomogram which are baseline T, baseline N, tumor length, induction chemo, lymph node metastasis by post chomoradiation PET and CT.

Figure3

Nomogram to assess the possibility of +ypNodes after preoperative chemoradiaton in patients with oesphageal cancer undergoing surgery.