Table 2 Identified descriptors and background factors for maximal lung cancer prediction performance.

From: Early symptoms and sensations as predictors of lung cancer: a machine learning multivariate model

BACKGROUND

Current smoking

Confirmed history of COPD

A cold, flu or pneumonia within the past 2 years

Confirmed history of pneumonia*

Female sex

Antibiotics within the past 2 years*

Older age (+1 SD, unit-variance scaled age)

BREATHING

5: Wheezing/panting*

30: Breathing worse when I lay down*

7: Gasped for breath

31: Breathing worse due to high humidity

12: Felt thickness in throat

33: Breathing worse due to coldness*

21: Breathing sound: Whistled

35: Breathing worse during certain times of the day*

29: Breathing worse upon exertion

COUGH

3: Sudden, loud cough*

11: Needed to clear my throat*

4: Hacking cough*

29: Cough varied over the day

5: Wheezing cough*

35: Cough varied over the year

6: Irritating, dry cough

63: Cough occurred/worsened when I exerted myself*

7: Coughed until I lost my breath, choked and/or vomited*

64: Cough occurred/worsened when I breathed deeply*

8: Cough attacks*

68: Cough worsened by high humidity

10: Small coughs*

PHLEGM/EXPECTORATES

3: Decreased amount*

24: Thin, fluid-like consistency*

6: White mucus or sputum*

25: Taffy-like/viscous consistency*

15: Haemoptysis/hematemesis (blood-mixed/brown sputum)

PAIN/ACHES/DISCOMFORT

3: Hurting: Comes and goes

67: Heartburn

8: Aches: Consistent

201: Pain/aches/discomfort: Throat*

9: Aches: Comes and goes

204: Pain/aches/discomfort: Shoulder blade

10_11_12: Aches: Positional/breathing-based

207: Pain/aches/discomfort: Shoulder(s)

14: Pain: Consistent

210: Pain/aches/discomfort: Neck

16_17_18: Pain: Positional/breathing-based*

213: Pain/aches/discomfort: Chest

27: Cramping aches/pains: Comes and goes*

223: Pain/aches/discomfort: Back

39: Dull aches/pain: Comes and goes

227: Pain/aches/discomfort: Moves around*

49: Tenderness

FATIGUE

VOICE CHANGES

3: Less strength, got weaker

1: Voice got more hoarse

4: Legs cannot cope

2: Voice got more rough/coarse

11: Felt constant tiredness, weakness, or lack of energy*

6: Cleared my throat more when I talked*

APPETITE/EATING/TASTE CHANGES

OLFACTORY CHANGES

1: Appetite loss

1: More difficult to distinguish smells

2: Enjoyed food less than before

2: Lost sense of smell*

5: Early satiety (feeling full quicker)

3: Heightened sensitivity to different smells

FEVER

OTHER CHANGES

1: Chills*

1: Cramps in calves

4: Felt cold

10: Drier skin*

13: Night sweats

13: Drier mouth

 

19: Feeling unfit

  1. Variables included in the final model (n = 70) are shown, including 7 background variables and 63 descriptors. Numbers indicate the identifiers of each of the included descriptors for each respective module and serve as a key to the regression coefficients shown in Supplementary Fig. S3. Of originally 285 descriptors, 145 met inclusion criteria (at least 4 observations in each group, lung cancer or no cancer). Additionally-excluded descriptors (n = 82) and background variables (n = 9) for model finalisation are indicated in Supplementary Table S2. History of chronic obstructive pulmonary disease (COPD) and history of pneumonia, respectively, are physician-confirmed. Bolded descriptors reached significance in terms of regression coefficients and 95% jack-knifed confidence intervals (ordered by strength of association to lung cancer, see Supplementary Fig. S3).
  2. *Indicates variables that had an average regression coefficient with an inverse association to lung cancer (n = 28).
Back to article page