Undetected Pancreatic Adenocarcinoma on CT: Frequency According to CT Scan Protocol

Purpose: CT is a main diagnostic modality for detecting pancreatic adenocarcinoma. This study aims to assess the frequency of missed pancreatic adenocarcinoma on CT scans according to different CT protocols. Methods: Consecutive pancreatic adenocarcinoma patients were retrospectively collected (12/2011-12/2015). Patients with abdominal CT scans performed up-to a year prior to cancer diagnosis were included. Two radiologists registered in consensus the presence and radiological signs of missed cancers. The frequency of missed cancers was compared between portal and pancreatic/triphasic CT protocols. Results: Overall, 180 CT scans of pancreatic adenocarcinoma patients were retrieved. 126/180 (70.0%) were pancreatic/triphasic protocols and 54/180 (30.0%) were portal protocols. The overall frequency of missed cancers was 6/180 (3.3%). The frequency of missed cancers was higher in portal CT protocols compared to pancreatic/triphasic protocols: 5/54 (9.3%) vs. 1/126 (0.8%), p=0.01. CT signs of missed cancers included: 3 cases of small hypodense lesions, 2 cases with peri-pancreatic fat stranding, 1 case of dilated pancreatic duct with a cut-off sign. Conclusion: The frequency of missed pancreatic adenocarcinoma is higher on portal CT protocols. Physicians should consider the cancer miss rate on different CT protocols.

Abdominal computed tomography (CT) is a main diagnostic tool for various gastrointestinal complaints, with millions of scans performed worldwide every year.CT commonly serves as a primary imaging modality for the detection of pancreatic malignancy, and considered the gold-standard for determining staging.
Demonstration of pancreatic adenocarcinoma (PDAC) is optimal using a dedicated biphasic CT protocol, comprised of a pancreatic phase scan (starting 35-45 seconds after intravenous contrast iodine-based injection), and a portal phase (performing another scan 65-70 seconds from contrast material injection) [10][11][12][13][14][15].Pancreatic phase is optimal for PDAC detection, making this desmoplastic tumor conspicuous compared to the normally enhancing pancreatic parenchyma.Portal phase is superior for assessing regional and distant spreading to peritoneum and liver.
Survival from PDAC depends on early detection, with surgical resection being the only potentially curative therapy [3,5].Therefore, Clinicians must be aware of the reliability and potential pitfalls of different CT protocols when diagnosing pancreatic malignancy.
Despite the importance of avoiding delayed diagnosis, there are only a scant number of publications regarding the rate of missed pancreatic adenocarcinoma.
This study aims to assess the frequency of missed pancreatic adenocarcinoma on CT scans according to different CT protocols.

Study design
The Chaim Sheba Medical Center at Tel-Hashomer Hospital Institutional review board committee (IRB) approval was granted for this retrospective study.Informed consent was waived by The Chaim Sheba Medical Center at Tel-Hashomer Hospital Institutional review board committee.All research methods were performed in accordance with relevant guidelines and regulations in accordance with the Declaration of Helsinki.
Consecutive subjects with histopathological diagnosis of pancreatic adenocarcinoma were retrospectively retrieved using a computerized search in our department's Radiological Information System (RIS) (12/2011-12/2015).
Only subjects with a new diagnosis of PDAC were included.Other pathologies, such as pancreatic neuroendocrine tumors, were excluded.Demographic, clinical data and CT referral indications were retrieved from the electronic medical records.
For each patient, we collected the CT scan in which PDAC was rst radiologically observed.These scans were considered the gold standard in the study.
For each patient, we collected all CT scans performed up to a year prior the gold standard.For each scan, we checked whether signs of pancreatic cancer have been missed by the original reporting radiologist.All missed cases were re-evaluated in consensus by two senior radiologists.
Missed PDAC was de ned as either a pancreatic mass that was not reported, or as secondary radiological ndings of malignancy (peri-pancreatic fat stranding, dilatation of pancreatic duct with or without a cut-off sign and peri-pancreatic lymphadenopathy) that were not reported.Small tumors were de ned as having a diameter of less than 20 mm as in Yoon et al. paper [20].
The study cohort inclusion process is presented in Chart 1.

Imaging technique
CT scans were performed in several institutes.The scans were categorized according to the CT protocol that was used: 1) A dual phase pancreatic-protocol CT, which comprises a pancreatic phase performed with a scan delay of 35-45s following a bolus of intravenous contrast agent, and a portal venous phase, performed with a scan delay of 65-70 s. 2) A triphasic CT protocol with late arterial (scan delay 30-45 s) and portal phases (scan delay 65-70 s) followed by an additional delayed scan 3 -5 minutes following intra-venous (IV) contrast bolus.3) A portal CT protocol (scan delay 65-70 s).All protocols were conducted with noncontrast scan prior to injection of intra venous contrast.The various protocols examined in the study were summarized in Table 1.

Statistical analysis
Descriptive statistics was used to summarize the study's characteristics.All analyses were conducted with SPSS (Version 20 Armonk, NY, US).Statistical signi cance was established at a 2-sided P < .05.Differences in miss rates were compared between the portal CT protocol group and the pancreatic/triphasic CT protocols (Fisher's exact test).We also compared differences in miss rates between university hospitals and outpatient clinics and between portal CT scans and pancreatic/triphasic CT scans (Fisher's exact test).
Thus, 180 patients were included in the study.Table 2 summarizes the study cohort according to CT protocol.The age range was 12-90 years, as there was a single case of a child with Fanconi anemia related tumor.
The referral indications for performing CT scans were retrieved and were grouped into fourteen categories (Table 3).Patients could have more than one referral indication.The most frequent indication was abdominal pain (96/180 patients, 53.3%), followed by weight loss (63/180, 35.0%).The "other" category included lower limb thrombosis, ischemia or edema, heartburn, splenomegaly, elevated liver enzymes, shoulder, ank or chest pain.Most incidental masses were detected on CT surveillance for previous malignancy (breast, prostate and lymphoma) or pancreatic/liver cyst follow-up.Two incidental masses were found during a work-up for bleeding ulcer and trauma.
The following CT signs were retrospectively identi ed in missed cancers: Three patients demonstrated small (diameter ≤20 mm) hypodense lesions (Figure 1).Three additional patients did not reveal a detectable pancreatic mass, but exhibited the following radiological ndings: two cases of peri-pancreatic fat stranding (Figure 2) and one case of dilated pancreatic duct with a cut-off sign (Figure 3).

Discussion
In this study, we have evaluated the failure to diagnose PDAC in different CT protocols.Our cohort included 180 patients, and found a miss rate of 9.3% of PDAC in portal CT scans.To the authors' knowledge, this is the largest cohort to investigate missed PDAC in abdominal CT scans, with reference to different CT protocols.Kielar et al. described 13 errors that involved the pancreas out of 222 imaging errors, one of which was due to a missed pancreatic mass [16], and Donald et al. presented 558 diagnostic imaging errors, of which CT scans accounted for 43%, including 4 patients with missed pancreatic tumors [17].These studies, however, did not analyze the rate of missed pancreatic tumors solely, and did not separate results according to CT protocols.This study demonstrated signi cantly higher miss rate (9.3%) of PDAC in portal CT scans.Our ndings are supported by current literature stating that pancreatic phase is superior in demonstrating pancreatic adenocarcinoma, due to better tumor-to-pancreas contrast [10][11][12][13][14][15].In addition, the lower rate of missed pancreatic tumors on pancreatic and triphasic scans may also be attributable to radiologists being more conscious of subtle imaging ndings when interpreting pancreatic/triphasic CT scans, which usually hold a stronger relation to the clinical question of tumor detection [18].Since pancreatic adenocarcinoma is a lethal malignancy and early detection is the main chance of survival, physicians' knowledge of CT techniques is of crucial importance.
Several previous studies investigated the imaging ndings of insidious pancreatic tumors [20][21][22][23][24][25].Yoon et al. showed in their study of small (≤20 mm) PDAC, that approximately one-fourth of the small pancreatic masses were iso-attenuating.In that study, most of the small iso-attenuating tumors showed secondary signs; of which the most frequent were pancreatic duct abnormalities, including cut-off or dilatation [20].Ahn et al. demonstrated that focal hypo-attenuation and pancreatic duct dilation with or without interruption were the most useful ndings for avoiding delayed diagnosis of pancreatic cancer [21].
In our study, undetected tumors registered the following radiological signs: i) small (≤20 mm) hypodense lesions only retrospectively identi ed, ii) peri-pancreatic fat stranding and iii) dilated pancreatic duct with cut-off sign.We further recommend that these signs be emphasized in physician's training.
Errors in cancer diagnosis are likely the most harmful and expensive types of diagnostic errors [18][19].Physicians should take into account the appropriate CT technique when there is clinical suspicion of pancreatic malignancy.
Our study has several limitations.This is a retrospective study, but only a retrospective study can present an accurate estimation of the miss rate of PDAC.Secondly, CT scans included in our work were done in several institutes, which re ect real life variability.Thirdly, the number of missed cases is small, though statistical signi cance has been observed.
In conclusion, the frequency of missed PDAC is higher on portal CT protocols.Physicians should consider the cancer miss rate on different CT protocols.

Table 1 .
Different Computed protocols examined in cohort