Gastrointestinal Cancer

Survival Outcomes in Asymptomatic Patients With Normal Conventional

Imaging but Raised Carcinoembryonic Antigen Levels in Colorectal Cancer
Following Positron Emission Tomography-Computed Tomography Imaging
KHURUM KHAN,a,* AVANI ATHAUDA,a,* KATHARINE AITKEN,a DAVID CUNNINGHAM,a DAVID WATKINS,a NAUREEN STARLING,a GARY J. COOK,b
ELEFTHERIA KALAITZAKI,c IAN CHAU,a SHEELA RAOa
a
GI and Lymphoma Unit, Department of Medicine, bDepartment of Nuclear Medicine, and cDepartment of Statistics, The Royal Marsden
National Health Service Foundation Trust, London, United Kingdom
*Contributed equally.
Disclosures of potential conflicts of interest may be found at the end of this article.
Key Words. Colorectal cancer x FDG PET-CT x Carcinoembryonic antigen x Conventional investigations

ABSTRACT
Background. This study had two aims: (a) to evaluate the utility to detect recurrence was 49 of 56 (88%; 95% confidence
of fluorine 18-fluorodeoxyglucose (FDG) positron emission interval [CI], 76%–95%) and specificity was 28 of 32 (88%; 95%
tomography (PET)-computed tomography (CT) in detecting CI, 71%–97%). Twenty-seven of 49 (55%) patients with PET-CT-
occult disease recurrence with raised carcinoembryonic antigen detected relapsed disease were deemed eligible for further
(CEA)and(b)to establish theprognosticeffects ofearlydetection curative therapy; 19 (70%) went on to receive potentially
of disease recurrence in patients with colorectal cancer (CRC). curative therapy. The median time to progression (8.8 months
Patients and Methods. Clinico-pathological data were obtained [interquartile range (IQR), 4.5–19.1 months] vs. 2.2 months
from all consecutive patients undergoing CRC surveillance from [IQR, 0.7–5.6]), median overall survival (39.9 months [IQR,
2004 to 2010 who had an elevated CEA level (.3 ng/mL in 23.6–65.4 months] vs. 15.6 months [IQR, 7.3–25.7 months]),
nonsmokers, .5 ng/mL in smokers) but normal or equivocal and 5-year survival (36.8% [95% CI, 16.5%–57.5%] vs. 6.1%
conventional investigations. Histopathological confirmation or [95% CI, 1.1%–17.6%]; p # .001) were higher in patients who
a minimum of 12 months’ clinical and radiological follow-up received potentially curative therapy than in those who
were required to ascertain disease relapse. received noncurative therapy.
Results. A total of 1,200 patients were screened; of those, 88 Conclusion. FDG PET-CT is a highly sensitive and specific tool
(59% men; mean age, 66 years [SD, 9.6]) eligible patients (67 for the detection of occult CRC recurrence. In .50% of
with normal and 21 with equivocal results on conventional patients, recurrent disease may still be potentially amenable to
investigations) were identified. Recurrent disease was de- curative therapy. Long-term survival can be achieved in such
tected in 56 of 88 patients (64%).The sensitivity of FDG PET-CT patients. The Oncologist 2016;21:1502–1508

Implications for Practice: Colorectal cancer (CRC) patients who, on follow-up, have normal or equivocal results on clinical
investigations but raised carcinoembryonic antigen (CEA) levels pose a significant challenge to treating physicians. This
study supported the notion that the early use of fluorodeoxyglucose (FDG) positron emission tomography (PET)-computed
tomography (CT) may have predictive and prognostic value in management of such patients. Long-term disease control and
cure can be achieved in a subgroup of this patient population with low-volume disease relapse who are amenable to
potentially curative treatment strategies. Reassuringly, the sensitivity and specificity for recurrence did not significantly vary
as a function of the CEA level, suggesting that even with a minimal CEA rise, benefit can be attained by conducting FDG PET-CT
in a timely manner.

INTRODUCTION
Colorectal cancer (CRC) remains one of the most common may have favorable outcomes after surgical resection, nearly
cancers in the world, associated with a high morbidity and half will relapse with metastatic disease, most frequently
mortality [1, 2]. Although patients with early-stage disease during the first 3 years of follow-up [3–5]. Patients with
CME

Correspondence: Sheela Rao, M.D., F.R.C.P., Department of Medicine, GI and Lymphoma Unit, The Royal Marsden National Health Service Foundation
Trust, Downs Road, Sutton SM2 5PT, United Kingdom.Telephone: 44-020-8661-3156; E-Mail: sheela.rao@rmh.nhs.uk Received June 2, 2016; accepted
for publication July 21, 2016; published Online First on October 14, 2016. ©AlphaMed Press 1083-7159/2016/$20.00/0 http://dx.doi.org/10.1634/
theoncologist.2016-0222

The Oncologist 2016;21:1502–1508 www.TheOncologist.com ©AlphaMed Press 2016

Khan, Athauda, Aitken et al. 1503

synchronous or metachronous oligo-metastatic disease may therapy with curative intent for synchronous or metachronous
achieve long-term benefit from early detection and potential oligo-metastatic disease, including surgical resection, radio-
utility of curative options in their management [5]. frequency ablation (RFA), and radical chemoradiation, were
Although many prospective randomized controlled trials also deemed to be eligible.
comparing intensive to standard follow-up failed to demon- Surveillance strategy for patients with no metastatic
strate any effect on survival benefit [6–10], a more recent disease at our institute included follow-up every 3 months in
meta-analysis has suggested contradictory findings [11]. That year 1, follow-up every 6 months in years 2 and 3, and annual
study pooled the results of 11 randomized controlled trials follow-up for years 4 and 5, with CEA test performed on each
totalling 4,052 patients and suggested that intensive follow-up visit. Annual CTwas performed for the first 3 years, and routine
confers a survival advantage. Despite significant variation in colonoscopy was performed every 2–3 years. PET was not
what constituted "intensive" follow-up in the included trials, routinely performed in these patients. For patients who
the common denominator was the inclusion of regular underwent localized therapeutic options with curative intent
carcinoembryonic antigen (CEA) measurements and regular after being diagnosed with oligo-metastatic disease, surveil-
computed tomography (CT). lance strategy included follow-up with CEA every 3 months
A rise in CEA often precedes the development of symptoms during year 1, follow-up every 6 months during years 2–5, and
and is generally considered to be the first sign of CRC annual follow-up during years 6 and 7, with CEA performed on
recurrence [12]; regular CEA measurement is therefore a each visit. CT was done every 6 months for years 1 and 2 and
component of follow-up programs globally [12–17].There are, then annually during years 3–5. Colonoscopy was performed
however, some caveats to CEA use in clinical practice, including as per the routine follow-up scheme described earlier in this
the following: (a) CEA rise may vary according to site of relapse section. All patients were required to have a minimum follow-
and (b) CEA elevation may reflect false-positive results due up of 5 years. A data tool pro forma was designed and all
to a variety of benign and malignant conditions [17–19]. relevant information was retrieved from the electronic patient
Regardless of the cause, an elevated CEA level in a patient on record by the authors. The hospital’s Committee for Clinical
follow-up warrants further investigations. In the UK, CTof the Research approved the study.
chest, abdomen, and pelvis is generally considered the first
investigation for patients with raised CEA [20]. Magnetic CEA and Conventional Imaging Investigations
resonance imaging (MRI) is another useful imaging modality An elevated CEA level was defined as .3 ng/mL in nonsmokers
considered helpful in evaluation of possible pelvic recur- and .5 ng/mL in smokers according to the hospital laboratory
rences and in characterizing abnormalities seen within the guidelines. Because this was a single-center retrospective
liver on CT [16]. Although both techniques complement each study, all CEA measurements were performed within the
other, a subset of patients will not have evidence of disease hospital laboratory, which has clinical pathological accredita-
recurrence on CT and MRI despite an elevated CEA level. In tion. Conventional imaging investigations were defined as CT
such patients, the use of fluorine 18-fluorodeoxyglucose with or without MRI, ultrasonography, or colonoscopy. The
(FDG) positron emission tomography (PET)-CT has been time interval between CT and PET-CT is summarized in
explored in an attempt to detect occult recurrence not supplemental online Figure 1. The mean time interval was
visualized on CT and to help clarify indeterminate findings on 4.46 weeks (range, 1–11 weeks; median, 4 weeks).
CT [21]. Initial conventional investigation findings were reviewed
Despite the common belief among the oncology commu- at a multidisciplinary meeting before FDG PET-CT was
nity that aggressive management of oligo-metastatic dis- requested. The results of the conventional investigation
ease in CRC leads to better survival outcomes, significant workup were categorized as no recurrence or equivocal by
variations in what is meant by "intensive follow-up" exist in using the original investigation reports. Equivocal results were
clinical practice [5, 22]; moreover, in patients with raised CEA defined as reports with abnormalities of uncertain significance
and normal conventional investigations, there are limited for which it was unclear whether these represented recurrence
data to inform appropriate management. In the current study or benign findings. Patients were categorized according to the
we sought to (a) evaluate the utility of PET-CT in detecting conventional investigation report, aided by any additional
occult disease recurrence with raised CEA and (b) establish information provided by subsequent multidisciplinary meet-
the prognostic effects of early detection of disease ing (MDM) review of the conventional investigations (if such a
recurrence. meeting had taken place). The serum CEA level measured
closest to the time point of the FDG PET-CT scan was used for
MATERIALS AND METHODS analysis.

Methods FDG PET-CT

This retrospective study screened all patients with CRC FDG PET-CT studies were acquired by using a Gemini (Philips,
consecutively treated at the Royal Marsden (RM) National Amsterdam, The Netherlands, http://www.philips.com) PET-
Health Service Foundation Trust, UK, from April 2004 to CT scanner in accordance with the standard RM protocol.
September 2010. Patients with confirmed histological di- Patients were required to fast for a minimum of 4 hours before
CME

agnosis of CRC who had been on follow-up after curative the study. FDG (400 MBq) was injected intravenously if the
therapy and underwent FDG PET-CT for an investigation of an blood glucose level was ,10 mmol/L. Patients were rested for
elevated CEA after normal findings on conventional investi- 60 minutes before PET acquisition. Emission data were
gations were included. Patients who had already received acquired from the base of the skull to the upper thighs

www.TheOncologist.com ©AlphaMed Press 2016

 1504 PET-CT in Asymptomatic CRC Patients

(4 minutes per bed; average, nine acquisitions). Unenhanced Table 1. Baseline characteristics of patients and
CTwas performed from the base of the skull to the upper thighs previous treatment
for purposes of attenuation correction and image fusion for
Patient characteristic Value
anatomical localization (approximate milliampere-second, 50
per slice). Demographic
The original FDG PET-CT report was reviewed by one of the Male 52 (59)
investigators and categorized as follows: no recurrence, Mean age (SD), yr 66 (10)
unlikely recurrence, equivocal, likely recurrence, and definite Age range, yr 37–85
recurrence. Results categorized into likely or definite re- Primary site
currence groups were defined as a positive FDG PET-CT result,
Colon 61 (69)
whereas the remaining three groups were defined as negative
Rectum 25 (28)
FDG PET-CT results. Relapse was confirmed by histopatholog-
ical confirmation of recurrent disease or by a combination of Both 2 (2)
repeat imaging and clinical follow-up to assess for response or Histologic features
progression to treatment. Mucinous 15 (17)
Nonmucinous 73 (83)
Statistical Analysis Previous therapies
Sensitivity of FDG PET-CT in detecting recurrence was Chemotherapy 72 (82)
calculated by dividing the number of true recurrences Radiotherapy 16 (18)
detected with FDG PET-CT by the total number of patients
Treatment
with recurrence in the following 12 months after the scan;
Primary surgery only 71 (81)
specificity was calculated by dividing the number of true
negatives detected with FDG PET-CT by the total number of Liver metastasectomy 10 (11)
patients without recurrence in the following 12 months after Lung metastasectomy 4 (5)
the scan. The positive predictive value (PPV) of FDG PET-CT in Liver metastasectomy and liver RFA 1 (1)
detecting recurrence was calculated by dividing the number of Othersa 2 (2)
true recurrences detected with FDG PET-CT by the total Unless otherwise noted, values are expressed as number (percentage) of
number of patients with a positive FDG PET-CT result; negative patients.
a
predictive value (NPV) was calculated by dividing the number One patient was treated with resection of abdominal wall recurrence;
another patient with previous anastomotic recurrence was treated with
of true-negative findings detected with FDG PET-CT by the total neoadjuvant chemotherapy, chemoradiation, and surgery.
number of patients with a negative FDG PET-CT result. These Abbreviation: RFA, radiofrequency ablation.
parameters were calculated with 95% confidence intervals
(CIs) overall and according to CEA level (CEA levels of 4–6, 7–10,
and .10 ng/mL). To test associations between categorical the eligibility criteria were included in the study. Patients who
variables, chi-square or Fisher exact tests were used accord- underwent PET-CT for any other indications, such as pre-
ingly, whereas for continuous variables t-tests or nonparamet- exenteration, assessment of response to chemotherapy, or
ric equivalents were used. staging of liver metastases, or as part of staging investigations
The proportion of patients diagnosed with recurrence on for de novo or recurrent established metastatic disease were
FDG PET-CT who were eligible for further potentially curative excluded from the study.The mean age of patients was 66 years
treatment was calculated along with 95% CIs. Of patients who (SD, 10), and 59% were male. Of all eligible patients, 67 (76%)
received potentially curative treatment, time to disease and 21 (24%) had normal and equivocal findings on
progression (TTP) was calculated from the date of treatment conventional investigations, respectively. Most patients had
to date of first progression as detected by CT. This was also colon cancer (69%), and 72 patients had received previous
calculated for patients who did not receive potentially curative chemotherapy (82%). Baseline patient characteristics are
treatment for relapse detected by FDG PET-CT as time from summarized in Table 1.
noncurative treatment to date of first progression. Overall
survival (OS) was calculated for all patients as the time from Overall Sensitivity and Specificity
date of FDG PET-CT to death or last follow-up. Patients without Fifty-six of 88 (64%) patients were confirmed to have relapsed
an event were censored at date of last follow-up. disease within the 12 months after their FDG PET-CTscan. In all
Survival estimates and 95% CIs were determined by using patients proven to have relapse, PET detected the site of subtle
the Kaplan-Meier method. Stata software, version 13.1 (Stata relapse; the relapse was histologically confirmed as malig-
Corp., College Station,TX, http://www.stata.com) was used for nancy after biopsy or resection, or the lesions had progressed
the analysis. such that they became apparent on conventional imaging. FDG
PET-CT correctly identified recurrent disease in 49 of 56
RESULTS patients (88%). The median CEA was 11 ng/mL (interquartile
range [IQR], 7–16 ng/mL; range, 4–62 ng/mL) in the relapsed
CME

Patient and Tumor Characteristics group and 7 ng/mL (IQR, 69 ng/mL; range, 4–22 ng/mL) in the
A total of 1,200 patients who had a diagnosis of CRC and were group without relapse (p 5 .002). Thirty-two patients did not
on follow-up were screened; of these, 88 patients who relapse in the subsequent 12 months, of whom 28 were
underwent PET-CT because of any clinical indication and met identified as having a negative FDG PET-CT scan (88%). Four

©AlphaMed Press 2016 OTncologist

he ®
Khan, Athauda, Aitken et al. 1505

Table 2. Proportion of patients who relapsed and outcome of FDG PET-CT compared with the gold standard
Variable FDG PET-CT-positive, n (%) FDG PET-CT-negative, n (%) Total, n (%)
Recurrence within 12 months 49 (56) (true-positive) 7 (8) (false-positive) 56 (64)
No recurrence within 12 months 4 (5) (false-positive) 28 (32) (true-positive) 32 (36)
Total 53 (60) 35 (40) 88 (100)
Abbreviations: CT, computed tomography; FDG, fluorine 18-fluorodeoxyglucose; PET, positron emission tomography.

Table 3. Proportion of patients with recurrences according to CEA level and corresponding sensitivity and specificity of fluorine
18-fluorodeoxyglucose positron emission tomography-computed tomography in detecting recurrence
Sensitivity, Specificity, Proportion detected on PET eligible for
CEA level (ng/mL) Patients, n (%) Relapses, n (%) % (95% CI) % (95% CI) potentially curative treatment, n/n (%)
4–6 26 (30) 11 (42) 91 (59–100) 87 (60–98) 9/10 (90)
7–10 27 (31) 17 (63) 82 (57–96) 80 (44–98) 7/14 (50)
.10 35 (40) 28 (80) 89 (72–98) 100 (59–100) 11/25 (44)
Abbreviations: CEA, carcinoembryonic antigen; CI, confidence interval; PET, positron emission tomography.

patients were found to have a false-positive FDG PET-CT result therapy for the PET-detected disease recurrence, the median
and 7 patients a false-negative FDG PET-CT result. Overall number of further lines of therapy was 1 (IQR, 0–3); for those
sensitivity was 88% (95% CI, 76%–95%) and specificity was who did not receive curative therapy, the median number was
88% (95% CI, 71%–97%) for FDG PET-CT in the detection of 2 (IQR, 1–3).
recurrent disease after normal results on conventional
investigations. The PPV and NPV for FDG PET-CT to predict Survival Outcomes
recurrence were 93% (95% CI, 82%–98%) and 80% (95% CI, The median TTP for patients who received potentially curative
63%–92%), respectively. These results are summarized in therapy for the FDG PET-CT-detected recurrence was 8.8
Table 2. months (IQR, 4.5–19.1 months). In the patients not treated
with curative intent, the TTP was 2.2 months (IQR, 0.7–5.6
Sensitivity and Specificity According to CEA Level months). The median OS for the whole group of patients was
The mean CEA level was 12.2 (SD, 10) ng/mL, with a range of 31.6 months.The median OS was 39.9 months (IQR, 23.6–65.4
4–62 ng/mL and a median value of 9 (IQR, 6–15) ng/mL. months) for those who received potentially curative treatment
Patients were further categorized according to predefined CEA and 15.6 months (IQR, 7.3–25.7 months) for those who did not
levels: 4–6, 7–10, and .10 ng/mL. Although the proportion of (Figs. 1, 2). The 5-year survival rate in the curative group was
patients with relapsed disease increased with increasing CEA, 36.8% (95% CI, 16.5%–57.5%) and in the noncurative group,
the overall sensitivity and specificity for FDG PET-CT in 6.1% (95% CI, 1.1%–17.6%) (p # .001).
detection of recurrent disease were similar among the three Forty-four patients (80%) had died at last follow-up, 9
groups (Table 3). The proportion of patients eligible for cura- patients (16%) were still alive, and 1 patient was lost to follow-
tive therapy was significantly lower in the higher CEA groups up. The median duration of follow-up was 25 months.
(4–6 ng/mL, 90%; 7–10 ng/mL, 50%; .10 ng/mL, 44%; p 5 .04).
DISCUSSION
Eligibility for Further Curative Therapy FDG-PET is an established imaging tool with a relatively high
Most patients with potentially curative recurrence had sensitivity and specificity for detecting recurrence in patients
relapses confined to the liver or lung or had local pelvic with hematological and solid malignancies, including CRC
relapse (supplemental online Table 1). Sixteen percent of [23–26]; however, data on predictive values and survival
patients were demonstrated to have multiple sites of relapse outcomes of patients undergoing FDG-PET in combination
on FDG PET-CT. Fifty-five percent of patients with recurrent with CT (FDG PET-CT), in the presence of normal or equivocal
disease detected on FDG PET-CT were eligible for further findings on conventional investigations but raised CEA levels,
potentially curative therapy, as deemed by MDM discussion are lacking. To our knowledge, we present here the largest set
(27 of 49 patients; 95% CI, 41%–68%). Of these, 19 patients of data on diagnostic accuracy of FDG PET-CT in such patients;
went on to receive potentially curative treatment: Twelve moreover, survival outcomes of this patient group are also
received surgery alone (7 liver metastasectomy, 2 lung presented for the first time.
metastasectomy, 2 pelvic surgery, and 2 surgical excision of Although an individual series suggested that FDG PET-CT
isolated nodal/soft tissue recurrence). Three patients re- had high diagnostic accuracy, with sensitivity and specificity of
ceived RFA to liver lesions. The remaining 3 patients received 85.7% and 94.7% respectively, FDG PET-CT was used as a
CME

multimodality treatment: chemotherapy/chemoradiation/ primary imaging tool in that study when patients were found to
pelvic surgery, chemotherapy/RFA of the liver, and chemo- have raised CEA levels [27]. This study therefore had two
radiation followed by pelvic surgery (supplemental online major limitations: (a) The superiority of FDG PET-CT over
Table 2). For patients who received potentially curative conventional investigations could not be established, and (b)

www.TheOncologist.com ©AlphaMed Press 2016

 1506 PET-CT in Asymptomatic CRC Patients

Figure 1. Imaging results of a 68-year-old man who presented with a rising carcinoembryonic antigen level (from 2 to 4 ng/mL) 18 months
after surgery and adjuvant chemotherapy for a Duke’s C colon adenocarcinoma. (A): Contrast-enhanced CT demonstrated no evidence of
recurrence. (B): FDG PET-CT demonstrated an isolated liver lesion, subsequently confirmed on Tesla MRI (C, D). He proceeded to undergo
surgical resection of the liver metastasis.
Abbreviations: CT, computed tomography; FDG, fluorine 18-fluorodeoxyglucose; MRI, magnetic resonance imaging; PET, positron
emission tomography.

Figure 2. Imaging results of a 70-year-old woman who had previously undergone surgery alone for a carcinoma of the ascending colon and was
investigated fora carcinoembryonic antigen level that had increased from 2 to 9 ng/mL. (A): The CTscan was reported as normal. (B): FDG PET-CT
showed peritoneal FDG uptake within the pelvis, confirmed on pelvic MRI (C, D). With the guidance of FDG PET-CT, a subtle abnormality was
seen in retrospect at this site on the original CT scan (as shown by arrows). This recurrence was treated with palliative chemotherapy.
Abbreviations: CT, computed tomography; FDG, fluorine 18-fluorodeoxyglucose; MRI, magnetic resonance imaging; PET, positron
emission tomography.
CME

there was limited information on patients with occult disease levels. Moreover, in our series, equivocal FDG PET-CT
and normal results on conventional investigations. In contrast, results were counted as a negative so not to bias them in
our study represents the more clinically challenging group of favor of the performance of FDG PET-CT. However, if an
patients who require cautious interpretation of raised CEA equivocal result had instead been categorized as positive,

©AlphaMed Press 2016 OTncologist

he ®
Khan, Athauda, Aitken et al. 1507

or alternatively had been excluded from the analysis, the Finally, because of the varying time interval between the
sensitivity would have improved to 91%; this is more in conventional investigations and FDG PET-CT, it could be
keeping with the higher values reported in some of the PET suggested that there was a lead-time bias such that
case series [24, 28]. recurrences were more obvious on imaging by the time of
The false-positive FDG PET-CTresults observed in our series the FDG PET-CT, and that a repeat CT at this stage may also have
were in keeping with the known limitation of FDG PET-CT detected the recurrence. However, a subanalysis of the
(e.g., a case of reflux esophagitis mimicking an esophageal accuracy of FDG PET-CT according to the time interval elapsed
malignancy). In the false-negative group, one patient with a from conventional investigations found no significant differ-
subcentimeter lung nodule that was not avid for FDG on PET-CT ence in sensitivity in patients with a greater than 6-week
subsequently was confirmed to have metastasis after surgical interval between the two tests compared with those with a less
resection. Interestingly, in our series there was no significant than 6-week interval. In fact, surprisingly, the sensitivity for
difference in the sensitivity of FDG PET-CT in detecting detection of recurrence was lower in the group with a greater
recurrence in patients with mucinous histology (data not than 6-week interval between the two sets of investigations
shown) despite suggestions that it is less sensitive within this (66% vs. 93%). The very wide CIs in the .6-week group reflect
subgroup [29]. the relatively small number of patients within this group,
The unique aspect of our series is that we report on survival therefore making it difficult to draw any conclusions with
outcomes of patients in whom disease relapse was detected certainty.
after FDG PET-CT. The median OS for patients who received
potentially curative treatment was longer than for those who CONCLUSION
did not. These findings suggest that long-term disease control FDG PET-CT is a highly sensitive and specific tool for the
and/or remission can be achieved in a proportion of patients detection of occult colorectal cancer recurrence in patients
by detecting disease recurrence before it is evident with with a raised CEA and normal results on conventional
conventional imaging modalities. investigations. A significant proportion of these patients
One of the notable findings in our series was that the have early disease relapse amenable to radical therapy. Long-
sensitivity and specificity for recurrence did not signifi- term disease control or cure may be achieved in such
cantly vary as a function of the CEA level.The high sensitivity patients. Because the sensitivity of FDG PET-CT to detect
of FDG PET-CT for recurrence, even at low CEA levels, is of recurrence does not vary significantly with CEA level, it can be
particular clinical relevance to the appropriate investiga- used as a clinically useful test for recurrence as soon as the
tion of patients with marginally elevated CEA levels. Given CEA becomes abnormal. FDG PET-CT should therefore be
the negative predictive value of 80%, the current study considered in patients with an unexplained rise in CEA levels
reassures the clinician that if the FDG PET-CT result is without evidence of disease recurrence on conventional
negative, the likelihood of the patient relapsing in the investigations.
subsequent 12 months is very low. We have shown that 50%
of patients with a CEA of 4 ng/mL had recurrent disease, ACKNOWLEDGMENTS
demonstrating that the risk of relapse even with marginal We acknowledge the support provided by the National
elevation of CEA level can still be considered substantial. Institute for Health Research Biomedical Research Centre at
Furthermore, the proportion of patients eligible for cura- the Royal Marsden National Health Service Foundation Trust
tive therapy in the low CEA group was significantly higher, and the Institute of Cancer Research. Dr. Khurum Khan is
at 90%, in comparison with only 44% of those with a supported by the Robert McAlpine Charity.
CEA .10 ng/mL. This would be in keeping with the lower
burden of disease volume commonly observed in those with
AUTHOR CONTRIBUTIONS
lower CEA levels. Conception/Design: Khurum Khan, Avani Athauda, Katharine Aitken, David
Limitations are associated with the retrospective nature Cunningham, Naureen Starling, Gary J. Cook, Ian Chau, Sheela Rao
of this study. While looking at the proportion of patients Provision of study material or patients: Khurum Khan, Katharine Aitken, David
Cunningham, David Watkins, Sheela Rao
eligible for curative treatment may provide a surrogate Collection and/or assembly of data: Khurum Khan, Avani Athauda, Katharine
marker for survival parameters, it is not an accurate method Aitken
to determine long-term outcome and survival. Therefore, Data analysis and interpretation: Khurum Khan, Avani Athauda, Katharine
Aitken, David Cunningham, Eleftheria Kalaitzaki, Sheela Rao
confounding factors, such as presence of comorbidities, Manuscript writing: Khurum Khan, Avani Athauda, Katharine Aitken, Naureen
availability of expertise, and suitability for localized thera- Starling, Gary J. Cook, Ian Chau, Sheela Rao
Final approval of manuscript: Khurum Khan, Avani Athauda, Katharine Aitken,
peutic options, should be considered when our findings are David Cunningham, David Watkins, Naureen Starling, Gary J. Cook, Eleftheria
interpreted. However, bearing in mind the recruitment and Kalaitzaki, Ian Chau, Sheela Rao
ethical difficulties of conducting largescale prospective
clinical trials with comparator arms (no PET scan), our data DISCLOSURES
are extremely reassuring for clinicians adapting the approach David Cunningham: Amgen, AstraZeneca, Bayer, Celgene, Merrimack,
of active surveillance. The current study did not retrospec- MedImmune, Merck Serono, Sanofi (RF); Naureen Starling:
AstraZeneca (RF); Sheela Rao: Amgen, Celgene, Eli Lilly, Roche,
tively review the original conventional investigations or FDG Baxalta, Merck Serono (C/A). The other authors indicated no financial
CME

PET-CT reports; it was felt that this would introduce relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert
interobserver bias and thus would affect the credibility of testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/
the study. inventor/patent holder; (SAB) Scientific advisory board

www.TheOncologist.com ©AlphaMed Press 2016

 1508 PET-CT in Asymptomatic CRC Patients

REFERENCES
1. Siegel R, Naishadham D, Jemal A. Cancer 12. McCall JL, Black RB, Rich CA et al. The value 21. Xia Q, Liu J, Wu C et al. Prognostic significance
statistics, 2013. CA Cancer J Clin 2013;63:11–30. of serum carcinoembryonic antigen in predicting of (18)FDG PET/CT in colorectal cancer patients with
2. DeSantis CE, Lin CC, Mariotto AB et al. Cancer recurrent disease following curative resection of liver metastases: A meta-analysis. Cancer Imaging
treatment and survivorship statistics, 2014. CA colorectal cancer. Dis Colon Rectum 1994;37: 2015;15:19.
Cancer J Clin 2014;64:252–271. 875–881.
22. Mella J, Datta SN, Biffin A et al. Surgeons’
3. Folprecht G, Gruenberger T, Bechstein WO et al. 13. Baca B, Beart RW Jr., Etzioni DA. Surveillance follow-up practice after resection of colorectal
Tumour response and secondary resectability of colo- after colorectal cancer resection: a systematic cancer. Ann R Coll Surg Engl 1997;79:206–209.
rectal liver metastases following neoadjuvant chemo- review. Dis Colon Rectum 2011;54:1036–1048.
23. Flanagan FL, Dehdashti F, Ogunbiyi OA et al.
therapy with cetuximab: The CELIM randomised phase
14. Chau I, Allen MJ, Cunningham D et al. The Utility of FDG-PET for investigating unexplained
2 trial. Lancet Oncol 2010;11:38–47.
value of routine serum carcino-embryonic antigen plasma CEA elevation in patients with colorectal
4. Gan S, Wilson K, Hollington P. Surveillance of measurement and computed tomography in the cancer. Ann Surg 1998;227:319–323.
patients following surgery with curative intent for surveillance of patients after adjuvant chemother-
colorectal cancer. World J Gastroenterol 2007;13: 24. Maldonado A, Sancho F, Cerdan J et al. 16.
apy for colorectal cancer. J Clin Oncol 2004;22:
3816–3823. FDG-PET in the detection of recurrence in co-
1420–1429.
lorectal cancer based on rising CEA level. Experi-
5. Khan K,Wale A, Brown G et al. Colorectal cancer 15. Pfister DG, Benson AB 3rd, Somerfield MR. ence in 72 patients. Clin Positron Imaging 2000;3:
with liver metastases: Neoadjuvant chemotherapy, Clinical practice. Surveillance strategies after cura- 170.
surgical resection first or palliation alone? World J tive treatment of colorectal cancer. N Engl J Med
Gastroenterol 2014;20:12391–12406. 25. Flamen P, Hoekstra OS, Homans F et al.
2004;350:2375–2382.
Unexplained rising carcinoembryonic antigen
6. Ohlsson B, Pålsson B. Follow-up after colorectal
16. Fiocchi F, Iotti V, Ligabue G et al. Role of (CEA) in the postoperative surveillance of colorectal
cancer surgery. Acta Oncol 2003;42:816–826.
carcinoembryonic antigen, magnetic resonance im- cancer: the utility of positron emission tomography
7. Mäkelä JT, Laitinen SO, Kairaluoma MI. Five-year aging, and positron emission tomography-computed (PET). Eur J Cancer 2001;37:862–869.
follow-up after radical surgery for colorectal cancer. tomography in the evaluation of patients with
Results of a prospective randomized trial. Arch Surg 26. Barrington SF, Mikhaeel NG, Kostakoglu L et al.
suspected local recurrence of colorectal cancer. Clin
1995;130:1062–1067. Role of imaging in the staging and response
Imaging 2011;35:266–273.
assessment of lymphoma: Consensus of the In-
8. Kjeldsen BJ, Kronborg O, Fenger C et al. A 17. Moertel CG, Fleming TR, Macdonald JS et al. An ternational Conference on Malignant Lymphomas
prospective randomized study of follow-up after evaluation of the carcinoembryonic antigen (CEA) Imaging Working Group. J Clin Oncol 2014;32:
radical surgery for colorectal cancer. Br J Surg 1997; test for monitoring patients with resected colon 3048–3058.
84:666–669. cancer. JAMA 1993;270:943–947.
27. Gade M, Kubik M, Fisker RV et al. Diagnostic
9. Schoemaker D, Black R, Giles L et al. Yearly
18. Abir F, Alva S, Longo WE et al. The post- value of (18)F-FDG PET/CT as first choice in the
colonoscopy, liver CT, and chest radiography do not
operative surveillance of patients with colon cancer detection of recurrent colorectal cancer due to
influence 5-year survival of colorectal cancer
and rectal cancer. Am J Surg 2006;192:100–108. rising CEA. Cancer Imaging 2015;15:11.
patients. Gastroenterology 1998;114:7–14.
10. Primrose JN, Perera R, Gray A et al. Effect of 3 to 19. Tan E, Gouvas N, Nicholls RJ et al. Diagnostic 28. Shen YY, Liang JA, Chen YK et al. Clinical impact
5 years of scheduled CEA and CT follow-up to detect precision of carcinoembryonic antigen in the de- of 18F-FDG-PET in the suspicion of recurrent
recurrence of colorectal cancer: The FACS random- tection of recurrence of colorectal cancer. Surg colorectal cancer based on asymptomatically ele-
ized clinical trial. JAMA 2014;311:263–270. Oncol 2009;18:15–24. vated serum level of carcinoembryonic antigen
(CEA) in Taiwan. Hepatogastroenterology 2006;53:
11. Pita-Fernandez S, Alhayek-Ai M, Gonzalez- 20. Maas M, Rutten IJ, Nelemans PJ et al. What is
348–350.
Martin C et al. Intensive follow-up strategies the most accurate whole-body imaging modality for
improve outcomes in nonmetastatic colorectal assessment of local and distant recurrent disease in 29. Berger KL, Nicholson SA, Dehdashti F et al. FDG
cancer patients after curative surgery: A systematic colorectal cancer? A meta-analysis: Imaging for PET evaluation of mucinous neoplasms: Correlation
review and meta-analysis. Ann Oncol 2015;26: recurrent colorectal cancer. Eur J Nucl Med Mol of FDG uptake with histopathologic features. AJR
644–656. Imaging 2011;38:1560–1571. Am J Roentgenol 2000;174:1005–1008.

See http://www.TheOncologist.com for supplemental material available online.

CME This article is available for continuing medical education credit at CME.TheOncologist.com.
CME

©AlphaMed Press 2016 OTncologist

he ®