Study reference

Study characteristics

Patient characteristics

Index test

(test of interest)

Reference test

Follow-up

Outcome measures and effect size

Comments

Chiu 2003

Observational,

part 1 retrospective (the past 5 years)

part 2 prospective (January – October 2000)

patients with NSCLC who underwent complete resection.

Part 1, N=30

Male: 80%

Age, median (range): 69 (43-82)

Postoperative

1 yr: 47%

2yr: 30%

5yr:23%

T1 N0 M0 n=7

T2 N0 M0 n=15

T2 N1 M0 n=2

T3 N0 M0 n=1

T1 N2 M0 n=1

T2 N2 M0 n=4

Part 2, N=43

Male: 91%

Age, median (range): 71/51-80

IA n=6

IB n=20

IIA n=0

IIB n=4

IIIA n=9

IIIB n=2

IV n=1

Low-dose computed tomography of chest(LDCT)

Part 1: surveillance procedures included physical

examination, chest radiography (frontal and lateral projections), LDCT, and contrast-enhanced standard-dose CT (SDCT) of chest.

Part 2: outpatient clinics every

3 months in the first 2 years and every 6 months for the next 3 years. Physical examination, sputum cytology, serum carcinoembryonic

antigen, chest radiography, and LDCT of chest were

performed at each scheduled visit.

contrast-enhanced standard-dose CT (SDCT) of chest

If suspected to have tumor recurrence: brain CT, abdominal sonography, fiberoptic bronchoscopy,

and whole body bone scans. Other diagnostic procedures

as indicated. Tissue or fluid specimens were obtained

for cytopathological diagnosis whenever possible.

Cut-off point(s):

Time between the index test en reference test:

Part 1: simultaneously

Part2: ?

For how many participants were no complete outcome data available?

-

Part 1:

SDCT detected 7 recurrent cases (23%)

LDCT detected 6 recurrent cases; Sensitivity (95%CI) =

0.86 (0.487 to 0.974)

chest radiographs detected

2 recurrent cases; Sensitivity = 0.29 (0.082 to 0.641)

Part 2

14/38 (37%) recurrent disease

LDCT detected 11 recurrent cases; Sensitivity (95%CI)= 0.786 (0.524 to 0.924)

chest radiographs detected

3 recurrent cases; Sensitivity = 0.214 (0.076 to 0.476)

Reference test not a gold standard

Index and reference tests not assessed independently / blinded

Crabtree 2015

Observational study, retrospective

USA

January 2000–April 2013

patients who underwent resection for stage I lung cancer

Excluded if alternated CT and CXR; after induction therapy, or incomplete resection

N=554

Male gender

CT: 101 (44%)

CXR: 121 (38%)

Age (mean ± SD)

CT: 65 ± 11

CXR: 66 ± 11

T2a tumor

CT: 94 (41%)

CXR: 103 (32%)

Adjuvant therapy

CT: 48 (21%)

CXR: 11 (3%)

All patients chest x-ray, two to four weeks after resection; thereafter 2 options:

1. routine postoperative surveillance CT scan (CT); with or without contrast; default was low-dose non-contrast CT;

baseline CT scan approx. 3

months after resection, with follow-up imaging every 6 months for the first 2–3 years; annually thereafter for at least 5 years; n=232

2. routine chest radiograph (CXR); n=322

Reference test was unclear.

invasive diagnostic procedure

only after positive test results

mean time from surgery to diagnosis of successive malignancy

CT: 1.93 years

CXR: 2.56 years

Median duration of follow-up up (years)

CT: 2.5 (range 0.3 – 9.9)

CXR: 3.5 (range 0.1 – 13.1)

Recurrence or new primary lung cancer was found in 24% (135/554)

rate of successive malignancy

CT group: 63/232 =27%

CXR group: 72/322 =22%

% false positive test results

CT group: 9/232= 4%

CXR group: 1/322= 0.3%

Overall survival (from date of resection for index lung cancer)

5-year survival rates

CT group: 67.8%

CXR group: 74.8%

median survival

CT group: 7.63 years

CXR group: not reached p=0.603

Cox proportional hazard model: modality of surveillance imaging not associated with survival (HR=0.988; 95%CI 0.629 – 1.551), and only age, Charlson comorbidity index, and type of resection

The imaging modality was chosen by the treating physician.

Chest x-rays tended to be utilized more in the early period of this study

no details on chest radiography

“no sign differences between the groups in most baseline characteristics, with the

exception of pathologic T stage of index tumor, type of surgical resection, and adjuvant therapy”

Dunlap 2012

Observational study prospective

USA

Between March 2005 and January 2010

patients treated with 3- to 5-fraction SBRT for stage I NSCLC

mean biologic equivalent dose received was 150 Gy (range, 78-180 Gy).

minimum follow-up of 12 months

n=80

median age 73 yrs (range, 54-87)

PET CT

Follow up with serial CT imaging every 3 months (free

breathing helical CT imaging)

PET CT was used as an adjunct test to evaluate progressive changes on CT scans that

were considered worrisome for local failure

Tumor progression based on CT imaging was confirmed by PET, as defined by 20% increase in the maximum SUV from the pretreatment value

Biopsy was attempted to confirm tumor recurrence, but most patients were deemed high risk for biopsy.

The median follow-up

time for the entire cohort from the end of treatment was 24 months (range, 12.0-36.0 months).

Recurrence: 10/80=12.5%

Tumor recurrence found in 35 of 44 patients using CT alone. Combined with PET, 10 of the 44 patients had tumor recurrence

5/10 recurrences confirmed by biopsy; remaining 5 patients high risk for biopsy; all demonstrate radiographic progressive disease on serial imaging

% false positive

PET: 0%

consecutive patients

Ebright 2013

Observational study, retrospective

USA

February 2009 – February 2012

35 patients consecutively treated

with SBRT for biopsy-proven early-stage NSCLC (T1a-T3, N0 M0)

Surveillance with PET CT

every 3-month PET/d

chest for the first 2 years.

All PET/CT studies were performed on a GE Discovery STE 16 PET/CT scanner

scanned from skull base to mid thigh and injected with an average of 11.0 mCi 8FDG.

PET images were uncoupled from the diagnostic CT im

ages, and he/she was blinded to the PET result

Pathologic confirmation of mediastinal and/or hilar lymph node recurrence was performed with cervical mediastinoscopy, EBUS, and/or endoscopic oesophageal

 ultrasound.

Median follow-up was 12.8 months

Time between index test and reference test not reported

Number of recurrences found

CT: none

PET CT: n=10 (29%)

% false positive test results

CT: –

Pet CT: 4/10=40%

One group of patients in whom surveillance with both PET CT and CT

Frank 1995

Observational study

USA

January 1990-Aoril 1994

Patients with biopsy proven lung cancer, underwent radiation therapy

Mean age 68.4 yrs

%male: 60%

N=20 (NSCLC: N=16)

Stage I, n=4

Stage II, n=5

Stage IIA: n=10

Stage IIIB: n=1

1. PET-CT, at ECAT scanner; DUR>3 treshold for tumor detection

2. CT, at standard commercial scanner, nonionic iv contrast

All insterpreted by 1 of 4 staff radiologists (experienced in chest CT)

Meta: well-defined, round, or

elliptical, soft tissue density greater than 1.5 cm in largest

diameter

If possible biopsy (only in case of PET-demonstrated recurrent disease)

CT and PET studies were compared for anatomic information versus. evidence of viable tumor. Concordant results were considered true positive or negative (whether or not biopsy confirmation.) Discordant results were resolved either by biopsy or subsequent behaviour on serial studies.

Number of biopsies: ?

PET at an average of 4.3 months postradiation therapy

Follow up: range 2.5 to 40 months

Detecting subclinical lung cancer recurrence

40 PET scans (18 patients); 35 CT scans (17 patients)

32 paired comparisons PET – CT, 1 or more discordant findings in 7 patients

PET

Sensitivity: 100%

Specificity: 89.3%

CT

Sensitivity: 67%

Specificity: 85%

Authors remark PET: some ROIs may have a DUR value in the range of 2-3. Such lesions diagnosed as malignant if examiners considered

Them to be so (qualitative visual analysis).

Gambazzi 2019

RCT

Switzerland

October 2011 to August 2014

Patients with 18F- FDG–PET–positive tumor, stage of I to III NSCLC and completed a curative-intent treatment

for NSCLC.

N=96

Median age (IQ range)

PET: 67 (61-74)

CT: 61 (56-70)

% male

PET: 66%

CT: 74%

Stage I-IV (stage IV: solitary brain metastases that were completely resected)

Surgery n= 60

Surgery+chemo: n=27

Radio+chemo: n=3

Other: n=6

1. PET-CT; A Biograph MCT

40 PET scanner (siemens)

N=50

2. CE-CT; Multidetector CT examinations of the chest performed using a 64 or a 320 detector-row CT scanner

N=46

imaging procedure performed

 at 6-month intervals.

In case of recurrence diagnostic workup at least one of the following procedures:

• nonscheduled PET-CT or CE-CT scan,
• non-scheduled brain CT, bronchoscopy,
• therapeutic-intent surgical
• intervention

Cancer recurrence confirmed

histologically unless clear evidence of metastatic

disease

A total of 130 surveillance PET-CTs and 138 surveillance

CE-CTs were performed

Suspicion recurrence

PET-CT: n=25/50 patients (50%)

CE-CT: n=22/46 patients (48%)

Confirmed recurrence: n=14 each group

Detecting recurrence

Sensitivity (95%CI)

PET-CT: 0.88 (0.62–0.98)

CE-CT: 0.93 (0.68–1.00)

Specificity (95%CI)

PET-CT: 0.62 (0.42–0.79)

CE-CT: 0.72 (0.53–0.87)

Subgroup patients after surgery:

Sensitivity

PET-CT: 0.87

CE-CT: 0.92

Specificity

PET-CT: 0.67

CE-CT: 0.81

Scheduled surveillance images and the results of

additional investigations performed due to a suspicious radiologic finding or symptoms of recurrence were discussed

 at weekly interdisciplinary board

Nakajima 2013

Observational study, retrospective

Japan

Between June 2006 and June 2009

Patients treated with SABR for stage I non-small cell lung cancer

maximum cumulative tumor diameter was < 5 cm and the number of tumors in lung 1-3

follow up min 6 months

N=59 (66 tumors)

%male: 66%

Median age 75yrs (45-92)

N=50 primary lung cancer; n=16 metastastic lung tumors

PET-CT

Dual-modality PET-CT imaging using Aquiduo

At least 1 PET-CT performed

137 FDG-PET/CT scans from the 59 study patients; 94 PET/CT examinations performed post-SABR because

clinical or imaging concerns for disease progression

Clinical courses and CT images, PET / CT findings, and laboratory data

local recurrence: a suspicious mass which showed continuous enlargement on two or more follow-up CT after FDG-PET/CT with or without elevation of tumor marker or appearance of new lesion

Biopsy to confirm local recurrence could not be performed except in 1 patient for their poor lung function; In 7 cases, tumor recurrence was pathologically

confirmed by salvage surgery

The post-SABR median follow-up duration was 29 months (range 7–52).

Recurrence: 16/59=27%; outcome of n=9 was uncertain

In 7/16 patient local failure: proven tumor recurrence (salvage lung resection); remaining 9 patients no pathological confirmation;

At 6–12 months after SABR,

as for local recurrence, the combination of intensity 2 and

mass-like FDG uptake obtained sensitivity of 100 %,

specificity 95 %

At 6-12 months after SABR,

Cutoff value of SUVmax = 4.5: sensitivity 88.9 %

specificity 94.5 %,

at 12 months after SABR, the two variables, the combination of intensity 2 and mass-like FDG uptake or SUVmax>4.5:

sensitivity 100 %

specificity 100 %

WELKE DATA ??

evaluated patterns

of FDG uptake and the maximum standardized uptake value (SUVmax) for differentiating tumor recurrence from radiation fibrosis after SABR.

Ohno 2017

Observational study, Prospective

Japan

April 2012 and March 2015

consecutive postoperative NSCLC patients

n=96

% male: 54%

Mean age 72 yrs

stage IA n=57

stage IB n=18

stage IIA n=10

stage IIB n=6

stage IIIA n=5

PET/CT,

septa-less 3D acquisition mode and a commercially available PET/CT scanner

routine radiological examinations, including whole-body CE-CT

All radiological studies were completed in random order every 6 or 12 months

All radiological examinations of a given patient were performed within 1 month

pathological and follow-up examinations

Recurrence of NSCLC determined based on results of routine radiological examinations, integrated FDG-PET/CT, whole body MRI, PET/MRI, and tumor markers, and diagnosed by means of physical examination and diagnostic imaging of lesions consistent with the criteria for recurrent NSCLC

When clinically feasible, diagnosis was pathologically confirmed

Recurrence: n=17/96 (18%)

Biopsies: n=5

Follw up: n=12

PET-CT

Sensitivity 1.0 (95%CI (0.816 to 1)

Specificity=0.81 (0.71 to 0.881)

% false negative 0%

% false positive 19%

Routine radiological exam.

Sens.: 0.706 (0.469 to 0.867)

Spec.: 0.873 (0.782 to 0.93)

% false negative 29%

% false positive 13%

work was financially and/or technically supported by Toshiba Medical Systems, Bayer Pharma,

and Eizai.

NB also data MRI (not PICO)

Onishi 2011

Observational study, Prospective

Japan

January 2004 and April 2009

121 consecutive NSCLC patients, who had undergone complete resection

Male: 66%

Mean age: 71 yrs

stage IA: n=57

stage IB: n= 33

stage IIA: n=3

stage IIB: n=15

stage IIIA: n=9

stage IIIB: n=4

whole-body integrated FDG-PET/CT, performed with septa-less 3D acquisition mode

and a commercially available PET/CT scanner; CT: no iodinated contrast media

standard radiological examinations (for distant metastasis)

all patients underwent a chest MDCT, abdominal MDCT, a neck MDCT performed with 16- and 64-detector CT scanners

All conventional radiological images interpreted with

a PACS system (ShadeQuest) all integrated FDG-PET/CT images with a workstation

(GENE Xeleris).

probability of the presence of recurrence evaluated with a 5-point scoring system:

1, definitely absent; 2, probably absent; 3, equivocal; 4, probably

present; and 5, definitely present

The presence of recurrent NSCLC determined based on:

standard radiological examinations, integrated FDG

PET/CT and assessment of tumor markers, and diagnosed by means

of physical examination and diagnostic imaging of lesions consistent with the criteria for recurrent NSCLC

When clinically feasible,

the diagnosis was histologically confirmed

Recurrent disease: n= 26 / 121 (21.5%)

Biopsies in 16 of 121 NSCLC patients

Using ROC curves, Optimal threshold values:

PET/CT: 3

standard methods: 2

The optimal cut-off value of SUVmax for improving the diagnostic capability of qualitatively assessed PET/CT was determined as 2.5.

Diagnostic accuracy (95%CI)

Standard radiological

Examinations

sens.: 0.731 (0.539 to 0.863)

spec.: 0.737 (0.64 to 0.815)

Qualitatively assessed

integrated FDG-PET/CT

sens.: 0.808 (0.621 to 0.915)

spec.: 0.768 (0.674 to 0.842)

Qualitatively and

quantitatively assessed integrated FDG-PET/CT

sens.:0.731 (0.539 to 0.863)

spec.: 0.874 (0.792 to 0.926)

FDG-PET/CT images / standard radiological

Examinations

were prospectively and independently interpreted;

readers PET/CT were blinded to the results of standard radiological

 examinations

Opoka 2013

Observational study, retrospective

Poland

January 2008 – January 2012

72 patients with previously treated NSCLC, admitted to hospital due to suspected disease recurrence; based on clinical picture, brochoscopic investigation, chest X-ray and/or CT

NSCLC stage I-IV

patients with recurrence:

Mean age 66.5yrs (range 48-86)

Men 70.5%

Patients without recurrence

Mean age 64.7 yrs (range 37-82)

Men: 20.8%

PET CT

GE discovery STE device (N=69 in centre A) of Siemens Biograph Truepoint64 device (n=3 in centre B)

Threshold SUV value for benign lesions ≤ 2.5

CT, spiral CT

Histopathological diagnosis in majority of patients based on surgical specimens and/ or samples transthoracic biopsy

N=5: bronchial biopsy specimen

N=2: bronchoalveolar lavage cytology

Interval between CT and PET-CT was 4 weeks

Recurrence

PET: 49/72=68%

Diagnostic accuracy (95%CI)

Sensitivity=0.938 (0.832 to 0.979)

Specificity=0.85 (0.675 to 0.941)

Takeda 2013

Observational study

Japan

Between 2005 and 2011

273 NSCLCs in 257 patients were treated with SBRT

Doses: 50 Gy and 40 Gy per 5 fractions for peripheral and central lesions, respectively

minimum follow-up posttreatment

FDG PET/CT was 2 months for patients

having no recurrence and was 6 months for patients with suspected

recurrence

T1- T4 tumors

FDG PET/CT, one year after SBRT

Additionally,18F-FDG PET/CT when recurrences were

suspected.

Different thresholds

214 18F-FDG PET/CT scans for 164 localized NSCLC tumors in 154 patients.

CT (non-contrast-enhanced CT with high resolution CT, unless adjacent to vessels), tri-monthly during first two years, subsequent at 4–6-month intervals

Local recurrence was diagnosed by pathological confirmation or

defined as an increase of >25% in the cross-sectional tumor size on successive CT scans at least three times over a 6-month period.

median follow-up period was 24.9 months (range: 6.3–72.1)

local recurrence: 17 tumors ; n=3

proved by biopsy;n= 14 diagnosed based on enlargements on successive CT images; in these 17 patients, 21 scans

For SUVmaxs on early and late images, optimal thresholds were identified as 3.2 and 4.2.

Diagnostic accuracy (95% CI) for

the optimal threshold on

-early images (SUVmax 3.2)

sens.: 1.00 (1.00–1.00)

spec.: 0.96 (0.93–0.99)

-late images (SUVmax 4.2)

sens.:1.00 (1.00–1.00)

spec.: 0.98 (0.96–1.00)

Takenaka 2010

Observational study, prospective

Japan

Patients with NSCLC, complete resection

N=92

% male:71%

Mean age 71 yrs

Stage IA, n=41

Stage IB n=30

Stage IIA n=4

Stage IIB n=7

Stage IIIA n=10

1. whole-body integrated FDGPET/CT, performed with the septa-less 3D acquisition mode and a commercially available PET/CT scanner

(Discovery ST; GE Health Care)

CT: no iodinated contrast

media

2. standard radiologic examinations: brain MR imaging (using contrast), contrast-enhanced whole-body CT and bone scintigraphy; By specialised radiologists with >10 yrs experience

every 6 months

PET-CT / standard radiology: probability of the presence of recurrence on 5-point score (2 raters): 1, definitely absent; 5, definitely present

threshold value for the visual scoring system was set at 4.

presence of recurrent NSCLC: based on standard radiologic examinations, integrated FDGPET/CT and assessment of tumor markers, and diagnosed by means of physical examination and diagnostic imaging of lesions consistent

with the criteria for recurrent NSCLC; histological conformation when feasible.

> 24 months after treatment

Biopsies performed for 16/ 80 NSCLC

Patients; others diagnosed by follow-up examinations

Recurrence: n=12 (13%)

Integrated FDG-PET/CT

Sensitivity 0.82 (95%CI 0.523 to 0.949)

Specificity 0.89 (95%CI 0.802 to 0.94)

Standard radiological examinations

Sensitivity 0.73 (95%CI 0.434 to 0.903)

Specificity 0.91 (95%CI 0.832 to 0.958)

Consecutive patients

FDG-PET/CT and standard radiological examination of

all patients were performed within 1 month (mean: 15.6 days;range: 6–28 days).

PET-CT images available for radiologists in standard examinations

Two chest radiologists and two nuclear medicine physicians not involved in final diagnosis recurrence, analysed standard radiological and PET-CT examinations

Study reference

Study characteristics

Patient characteristics ²

Intervention (I)

Comparison / control (C) ³

Follow-up

Outcome measures and effect size ⁴

Comments

Karzijn 2016

Observational study, retrospective review of medical records

The Netherlands, one hospital

Between 2008 and 2014

NSCLC patients stage I or II, treatment consisted of either

SBRT or surgical resection

N=228 patients treated; 86 patients developed recurrence;

73 patients included in analysis

Treatment

SBBRT: n=22

Surgery: n=51

%male: 55%

Mean age 66.4yrs (sd 10.5)

Stage IA n=37

Stage IB n=17

Stage IIa n=14

Stage IIb n=5

Surveillance with receiving CT scans with or without chest radiographs (CT group)

N=23

Follow up: clinical visit every 3 months during the first 2 years and every 3 to 6 months the years thereafter

Surveillance with only chest radiographs (CR group)

N=50

-

Overall survival, median (IQR), months

I: 27.2 (18.5-53.2)

C: 22.1 (14.2-39.2)

HR=1.427, 95% CI 0.755-2.695, p=0.273) (corrected for ECOG and morphology)

Only patients who developed recurrence were included

Recurrence in surgical patients was confirmed either by biopsy or by 18F-FDG-PET scan; Recurrence in SBRT patients was determined by biopsy, by 18F-FDG-PET scan or determined as clinically based on imaging.

Authors conclusion “Follow-up with a chest radiography did not result in inferior clinical outcomes compared to follow-up with a CT scan”

Reddy 2017

Observational study, retrospective

USA, single tertiary cancer center

January 2000 to January 2011

patients with pathologically confirmed NSCLC stage III, treated with definitive radiation therapy to a dose of ≥ 60 Gy

at least 2 surveillance scans

no events / death < 6 months after treatment

n=200

median age 65 yrs

91.5% concurrent chemo

-PET: 29.5%

-CT:52.5%

Adjuvant chemo

-PET: 37.1%

-CT: 16.8%

Stage IIIA n=94

Stage IIIB n=106

PET/CT

Alternating PET/CT and CT scans for 1–2 years (since it is the biggest risk period), reserving PET/CT

for abnormal findings on CT scan.

If a PET/CT was performed during one of the initial two scans, the patient was categorized into the PET/CT-group

N=105

CT

CT scan at every surveillance visit, every 3–6months for the first 5 years and annually thereafter

patients that only received CTs as their initial two scans were categorized into the CT-group

n=95

Median follow-up for alive patients was 59.4 months (range, 14.5–125.3).

Recurrence in n=102 (51%)

PET:n=50 (47.6%)

CT:n=52 (54.7%)

Overall survival, median (months)

CT: 41.2

PET: 41.3

(no data on uncertainty)

multivariate analysis (corrected for KPS, stage IIIB, squamous histology, concurrent chemo): PET/CT (versus CT) not associated with improved overall survival (HR: 1.2, 95% CI: 0.83–1.7)

the ratio of CT to PET/CT dramatically increases over time reflecting the practice of omitting routine PET/CT in those patients who do not have local or distant failure in the initial 2 years