HAL Model of O'Connell et al for Prediction of Lymph Node Metastasis in Lung Cancer

Description

O'Connell et al developed a model for evaluating a patient with treatment-naive non-small cell lung cancer (NSCLC) with mediastinal lymphadenopathy. The authors are from M.D. Anderson Cancer Center, Cleveland Clinic, Henry Ford Hospital, and the Johns Hopkins University.

HAL = model to help with the assessment of adenopathy in lung cancer

Patient selection: treatment-naive non-small cell lung cancer (NSCLC)with lymphadenopathy, T1 to T3, N0 to N3, M0

Exclusion: small cell lung cancer, recurrent cancer, synchronous primaries

Outcome: lymph node metastasis by EBUS-TBNA, with N0 or N1 versus N2 or N3

Parameters for multivariate logistic regression:

(1) age in years

(2) location in the lung

(3) histology

(4) N stage by CT

(5) N stage by PET

Parameter	Finding	Points
age in years		(-0.0311 * (age))
location	outer two-thirds of lung	0
	central one-third of lung	0.5881
histology	adenocarcinoma	0
	squamous cell	-0.8235
	non-small cell carcinoma	0.1423
	other primary lung	-0.5079

PET	CT	Points
N0	N0 or N1	0
N0	N2 or N3	0.9694
N1	N0 or N1	1.5766
N1	N2 or N3	0.9009
N2 or N3	N0 or N1	2.3726
N2 or N3	N2 or N3	3.7531

X =

= SUM(points for all 5 parameters) - 1.0663

probability of

= 1 / (1 + EXP((-1) * X))

Performance:

• The reported area under the ROC curve was 0.85.

To read more or access our algorithms and calculators, please log in or register.

HAL Model of O'Connell et al for Prediction of Lymph Node Metastasis in Lung Cancer

Description

Specialty