Table 7 Research variable prediction

[121]

ANN

BPE/CaP

Analysis of variables of quality of life questionnaire

Questionnaire suggested by medical and allied professional

High- or low-quality group

Single centre recruitment with BPE or CaP, 63 cases

ROC, Linear quadratic and logistic regression

Ac 90%, Se 94%, Sp 85%, PPV 89%, NPV 92%

Identify relevant variables

[78]

ANN

Nlt

Stone recurrence after ESWL

Anatomy, position, stone analysis, urine analysis, previous stone, medical treatment

Stone recurrence

65 patients post ESWL from single centre

33 test set

ROC AUC vs LR

AUC 0.96, Se 91%, S 91%

Stone recurrence, fragments not risk factor

[122]

ANN

CaP

Biochemical failure post RRP

TNM, tPSA, Gleason, pathology stage

BCF at 3 years

Yes or no

564 patients’ data post RRRP with Gl 7, single centre

ROC, Kaplan Meier and Cox Proportional Hazards Model

AUC 75%, NPV 84

Gleason 7 is inversely correlated to disease free survival and direct to BCF

[122]

ANN

CaP

Biochemical failure post RRP

TNM, tPSA, Gleason, pathology stage

BCF post RRRP

564 patients’ data post RRRP with clinically localised CaP Gl7, single centre

ROC, Kaplan Meier for survival and Cox Proportional Hazards

AUC 81%, NPV 93%

[75]

ANN

Nlt

lower pole stone ESWL

Gender, BMI, radiology, stone size, composition, urine analysis, 24 h urine, serum ca and creatinine

Clearance or intervention

321 patients with lower pole stone

211 random set

ROC, Sp, Se, vs LR

AUC 0.97, Se 95%, Sp 92%

BMI, normal urinary transport and infundibular width of 5 mm or more and the infundibular ureteropelvic angle is 45° or more are correlated with stone clearance

[103]

FNM

Bca

Recurrence classifier

Age, gender, grade, smoking, previous cancer, p53, hMLH1, hMSH2

Recurrence or not

109 patients from one centre with bladder TCC

tenfold CV ROC, LR

AUC 0.98, Se 90%, Sp 80%, PPV 92%, NPV 74%, Ac 88%

p value calculated to compare all models, the effect of combining HK p53 with other variables

[103]

FNM

Bca

Survival predictor

Age, gender, grade, smoking, previous cancer, p53, hMLH1, hMSH2

Survival in months

109 patients from one centre with bladder TCC

tenfold CV

Root mean square

RMS = 4.8

[103]

ANN

Bca

Recurrence classifier

Age, gender, grade, smoking, previous cancer, p53, hMLH1, hMSH2

Recurrence or not

109 patients from one centre with bladder TCC

ROC, LR

10% cross validation

AUC 0.91, Se 94%, Sp 96%, PPV 99%, NPV 84%, Ac 95%

[103]

ANN

Bca

Survival predictor

Age, gender, grade, smoking, previous cancer, p53, hMLH1, hMSH2

Survival in months

109 patients from one centre with bladder TCC

10% cross validation RMS

RMS = 11.7

[123]

ANN

Bca

diagnosis

Urine levels of nuclear matrix protein-22, monocyte chemoattractant protein-1 and urinary intercellular adhesion molecule-2

Cancer and benign

253 Data from one cystoscopy clinic

ROC, Sp, Se

Se 100%, Sp 75.7%, PPV 32.9%, NPV 100%,

The three factors improve diagnosis

[124]

ANN

BPE

Significant LUT symptoms in BPE

Age, PSA, Qmax, TZV, TPV, Oss, ISS, PVR

Progression or no

397 patient with mild LUTS from 4 centres

1/3 CV

ROC, Sp, Se, Then sensitivity analysis

Ac 79%, Se 82%, Sp 77%, PPV 78%, NPV 81%

PSA, Oss, TZV are correlated to disease progression

[125]

ANN

Hgon

Diagnosis of hypogonadism,

Age, ED, depression score, sexual health score, testosterone level

Risk of hypogonadism

148 one centre

70 test cases

Depression most significant, p < 0.0019

[126]

ANN

BPE/CaP

Diagnosis of BPE and CaP

Age, tPSA, %f PSA, TPV, MIC-1, Hk11, MIF

Cancer and benign

Single centre 371 patients

LOO

AUC 0.91, Se 90%, Sp 80%

Positive if all makers added together

[127]

ANN

Bca

Survival and recurrence predictor

22 different genes variables

Risk and time to relapse

67 bladder neoplasms and 8 normal bladder specimens

Difference RMS

10 folds CV ROC AUC

RMS 5.2

Ac 100%

500 genes where reduced to 22 genes for creating the network, thus significant

[127]

FNM

Bca

Survival and recurrence predictor

66 rules from 11 gene variables

Risk and time to relapse

67 bladder neoplasms and 8 normal bladder specimens

Difference RMS

10 folds CV ROC AUC

RMS 2.2

Ac 100%

500 genes where reduced to 22 genes for creating the network, thus significant

[105]

FNM

Bca

Recurrence (classifier)

Age, gender, grade, smoking, previous cancer, p53, methylation index (% of loci on chromosomes), RARB methylation

Recurrence or not

117 patients with 1ry TCC or UCC from one centre

10% cross validation ROC, LR

AUC 0.98, Se 88–100%, Sp 94–100%, Ac 100%

p value calculated to compare all models, the effect of combining HK p53 with other variables

[105]

FNM

Bca

Survival predictor

Age, gender, grade, smoking, previous cancer, p53, methylation index (% of loci on chromosomes), RARB methylation

Survival in months

117 patients with 1ry TCC or UCC from one centre

10% CV

Kaplan Maier for survival

Average error = 5 months

Interrogate different markers to suggest a predicative combination

[105]

ANN

Bca

Recurrence (classifier)

Age, gender, grade, smoking, previous cancer, p53, methylation index (% of loci on chromosomes), RARB methylation

Recurrence or not

117 patients with 1ry TCC or UCC from one centre

10% cross validation ROC, LR

Ac 89–90%, Se 81–87%, Sp 95–100%

[105]

ANN

Bca

Survival predictor

Age, gender, grade, smoking, previous cancer, p53, methylation index (% of loci on chromosomes), RARB methylation

Survival in months

117 patients with 1ry TCC or UCC from one centre

10% CV

Kaplan Maier for survival

Average error = 9 months

p for comparison ANN and FNM calculated

[128]

ANN

CaP

Diagnosis of cancer in PSA 1–4 4–10

Age, tPSA, %fPSA, TPV, DRE, -5pro PSA, -7, pro PSA

Risk of cancer

2 centre PSA 1–10 and TRUS 6–12 cores, 898 patients

ROC, Spearman correlation co efficient LOO

AUC 84%

Pro PSA improved detection rate in 1–4 and improved %fPSA performance in 4–10 group

[129]

ANN

CaP

Early CaP diagnosis

Age, tPSA, %fPSA, hK11, hK11/tPSA, hK11/%tPSA

Cancer or benign

357 with histologically proven cancer or BPE

ROC Se, Sp test set 206 with histologically proven cancer or BPE

AUC 0.84, Se 90%, Sp 52%

Sensitivity analysis of these variables to demonstrate their impact on AUC

[130]

ANN

CaP

Early CaP diagnosis

Age, tPSA, %f PSA, TPV, DRE (PSA done by five different assays)

Risk of cancer

585 patients with suspected cancer PSA 0.49–27

ROC AUC

25% random set 195 patients and LOO

AUC 0.91 (mean value)

Authors suggests developing PSA assay specific ANN to optimise function

[131]

ANN

CaP

Prostate cancer early diagnosis

Age, BMI, tPSA, fPSA, TPV, PSAD, smoking, systolic-diastolic pressure, pulse, Gl

Cancer or benign

300 patients’ data with suspected cancer from one centre

10- folds CV

ROC Se, Sp

Ac 79%, Se 81%, Sp 78%

[131]

SVM

CaP

Prostate cancer early diagnosis

Age, BMI, tPSA, fPSA, TPV, PSAD, smoking, systolic-diastolic pressure, pulse, Gl

Cancer or benign

300 patients’ data with suspected cancer from one centre

10- folds CV

ROC Se, Sp

Ac 81%, Se 84%, Sp 75%

Smoking is a significant classifier but not BMI

[132]

ANN

CaP

Diagnosis

Age, tPSA, %f PSA, DRE, TPV

Risk of cancer

PSA2-20 393proscpective data

ROC AUC

LOO

AUC 0.75, Se 90%, Sp 37%

Demonstrate the impact of different data cohorts on ANN performance

[133]

FNM

ANN

Bca

Gene micro array to predict UCC progression

200 genes reduced from 2800 by Pearson correlation

Cancer progression to muscle invasive or metastatic

66 tumours from 34 patients in one centre

COX multivariate analysis

10 folds CV

11 new gene signatures

200 gene micro array reduced to 11 gene signatures

[134]

ANN

U Dyn

Urodynamic interpretation

Age, BMI, menopause, sexual activity, UTI, number of vaginal deliveries, surgery,

U Dyn diagnosis

802 data from single centre POP with symptoms and UDS performed

ROC and compare to multi linear regression CV 20%

AUC 80% (Average)

ANN cannot replace Urodynamic

[135]

ANN

Fert

Seminal profile from questionnaire about life habits and health status

Age, BMI, demographic, medical history facts, smoking, alcohol, life style and cloth

Seminal profile

100 volunteers one centre study

ROC AUC Se, 10 Folds cross validation

Se 73–94%, Sp 25–45%, PPV 79–92%, NPV 7.4–54%

Comparison of different AI classifiers with same variables

[135]

SVM

Fert

Seminal profile from questionnaire about life habits and health status

Age, BMI, demographic, medical history facts, smoking, alcohol, life style and cloth

Seminal profile

100 volunteers one centre study

ROC AUC Se, tenfold CV

Se 74_99%, Sp 12–21%, PPV 75–91%, NPV 4–86%

[135]

DT

Fert

Seminal profile from questionnaire about life habits and health status

Age, BMI, demographic, medical history facts, smoking, alcohol, life style and cloth

Seminal profile

100 volunteers one centre study

ROC AUC Se, tenfold CV

Se 72–96%, Sp 12–41%, PPV 77–90%, NPV 4–48%

[136]

ANN

Fert

Seminal profile from questionnaire about life habits and health status

Age, season, childhood disease, surgery, trauma, smoking, alcohol, hours sitting ANNA1

Sperm concentration

100 volunteers one centre study

ROC AUC Se, Sp

10 Folds CV

Se 95%, Sp 50%, PPV 93%, NPV 60%

[136]

ANN

Fert

Seminal profile from questionnaire about life habits and health status

Age, BMI, marital status, vaccines, siblings, allergy, baths, hours of sleep ANNA2

Sperm motility

100 volunteers one centre study

ROC AUC Se, Sp

Se 89%, Sp 44%, PPV 89%, NPV 44%

[137]

ANN

CaP

Statistical evaluation of PSA INDEX

Age, TPV, DRE, tPSA, %fPSA

Risk of Cancer

1362 from multiple centres with suspected CaP and PSA 1.6–8.0

ROC AUC and comparison to other markers

AUC 0.7—0.74

[137]

ANN

CaP

Statistical evaluation of PSA INDEX

Age, TPV, DRE, tPSA, %fPSA, %p2PSA

Risk of Cancer

1362 from multiple centres with suspected CaP and PSA 1.6–8.0

ROC AUC and comparison to other markers

AUC 0.73—0.79

[137]

ANN

CaP

Statistical evaluation of PSA INDEX

Age, TPV, DRE, tPSA, %fPSA, %fPSA prostate health index (p2PSA / fPSA X square root tPSA)

Risk of Cancer

1362 from multiple centres with suspected CaP and PSA 1.6–8.0

ROC AUC and comparison to other markers

AUC 0.73- 0.8

Prostate Health index improved ANN performance

[112]

ANN

Bca

Survival post cystectomy

Age, gender, albumin, surgical approach, tumour stage, follow up period, type of diversion

5 years survival

117 patients with post cystectomy from one centre

ROC, Se, Sp Ac, 10 Folds cross validation

Ac 72–80%

Comparison of 7 different machine learning

RELM and ELM had best performance

[138]

ANN

CaP

 + ve lymph nodes to the total number of lymph nodes in predicting BCF

Age, tPSA, Clinical stage, Gl, seminal vesicle invasion, number of positive lymph nodes and laterality of lymph node involvement

BCF

124 cases with lymph node dissection

hazard ration for each variable

LND, Gl, and stage were identified as independent prognostic

LND is more prognostic than their number

[139]

BN

BPE

Correlation between symptoms, decision and outcome of surgery

Age, Qmax, PVR, PSA, TPV, TZV, BOO on UDS, and IPSS scores (stratified)

surgical decision-BN model, the outcome of surgery

1108 cases from one centre

ROC AUC and correlation coefficient

AUC 0.8 TZV (R = 0.396, P < 0.001), treating physician (R = 0.340, P < 0.001) and BOO on UDS (R = 0.300, P < 0.001)

TPV, physician, BOO on UDS, and the IPSS item of intermittency were factors that directly influenced

Decision-making in physicians treating patients with LUTS/BPE

[140]

ANN

CaP

Progression biomarkers

Gene microarray

Cancer progression and DSS

192 tissue histology results

MSE for each variable, then Kaplan Meyer and Pearson’s × 2-tests

10 gene microarrays identified by ANN

Ki67 and DLX2, appear to predict CaP specific survival and metastasis

[141]

ANN

VUR

Renal ultrasound to predict voiding cystourethrogram (VCUG)

Renal ultrasound findings

abnormal VCUG

2259 cases post UTI and had VCUG

ROC AUC

Se 64.2%, Sp 59.6%, PPV 61.6%, NPV 62.2%, AUC 0.6852

Renal ultrasound is a poor screening test for VCUG-identified abnormalities