Table 3 Multivariable Cox analysis of disease-free survival in validation cohort 2

From: Whole slide image based deep learning refines prognosis and therapeutic response evaluation in lung adenocarcinoma

 

Validation cohort 2 (n = 175)

 

Univariable

Multivariablea

Multivariableb

Variables

HR (95% CI)

P

HR (95% CI)

P

HR (95% CI)

P

Age (continuous)

1.01 (0.99–1.05)

0.347

    

Sex (male vs. female)

0.92 (0.53–1.61)

0.777

    

Smoke (yes vs. no)

1.55 (0.84–2.88)

0.164

    

Location (left vs. right)

1.06 (0.60–1.88)

0.833

    

Surgery (lobectomy vs. sublobar resection)

1.40 (0.44–4.51)

0.571

    

EGFR mutation (positive vs. negative)

0.70 (0.39–1.24)

0.218

    

KRAS mutation (positive vs. negative)

1.75 (0.69–4.41)

0.237

    

VPI (present vs. absent)

3.91 (2.23–6.84)

<0.001

1.59 (0.86–2.95)

0.138

1.44 (0.79–2.62)

0.238

STAS (present vs. absent)

3.79 (2.13–6.75)

<0.001

1.30 (0.69–2.45)

0.421

1.15 (0.60–2.21)

0.666

LVI (present vs. absent)

3.90 (2.20–6.93)

<0.001

1.66 (0.90–3.05)

0.106

1.48 (0.80–2.74)

0.214

IASLC grade

      

 II vs. I

2.21 (0.23–21.23)

0.493

1.94 (0.20–18.67)

0.568

1.88 (0.19–18.13)

0.587

 III vs. I

22.22 (3.06–161.32)

0.002

9.02 (1.13–72.27)

0.038

10.29 (1.28–82.93)

0.028

TNM stage

      

 II vs. I

3.63 (1.64–8.04)

0.002

1.66 (0.71–3.87)

0.244

1.54 (0.67–3.56)

0.314

 III vs. I

7.14 (3.80–13.40)

<0.001

2.73 (1.36–5.51)

0.005

3.08 (1.53–6.21)

0.002

Adjuvant chemotherapy (yes vs. no)

1.78 (1.01–3.12)

0.046

0.76 (0.42–1.40)

0.385

0.69 (0.38–1.25)

0.217

WSI-based score (high risk vs. low risk)

3.04 (1.62–5.72)

0.001

  

2.96 (1.55–5.66)

0.001

  1. EGFR epidermal growth factor receptor, KRAS Kirsten rat sarcoma viral oncogene, VPI visceral pleural invasion, STAS spread through air space, LVI lymph-vascular invasion, IASLC International Association for the Study of Lung Cancer, WSI whole slide image, HR hazard ratio, CI confidence interval, DFS disease-free survival.
  2. aMultivariable model without incorporating the model defined low- and high-risk groups.
  3. bMultivariable model with incorporating the model defined low- and high-risk groups.