Better retrieval of lymph nodes in colorectal resection specimens by pathologists’ assistants
- C C H J Kuijpers1,
- H J van Slooten1,
- W H Schreurs2,
- G R H M Moormann2,
- M A Abtahi1,
- A Slappendel1,
- V Cliteur1,
- P J van Diest3,
- N M Jiwa1
- 1Symbiant Pathology Expert Centre, Alkmaar, The Netherlands
- 2Department of Surgery, Medical Centre Alkmaar, Alkmaar, The Netherlands
- 3Department of Pathology, University Medical Center Utrecht, Utrecht, The Netherlands
- Correspondence to Dr N M Jiwa, Symbiant Pathology Expert Centre, Alkmaar, The Netherlands;
- Received 12 September 2012
- Revised 12 September 2012
- Accepted 12 September 2012
- Published Online First 19 October 2012
Background Errors in surgical pathology are partly due to the increasing workload of pathologists. To reduce this workload, ‘pathologists’ assistants’ (PAs) have been trained to take over some of the pathologists’ recurrent tasks. One of these tasks is the precise examination of ≥10 lymph nodes (LNs), which is of paramount importance to reduce the risk of understaging of colorectal cancer patients.
Aims To evaluate the role of PAs in harvesting LNs in colorectal resection specimens and, by doing so, in improving patient safety.
Methods LN harvest was retrospectively reviewed in 557 pathology reports on colorectal resection specimens collected in two Dutch hospitals from 2008 until 2011.
Results PAs sampled ≥10 LNs in significantly more cases than pathologists did (83.2% vs 60.9% in hospital A and 79.2% vs 67.6% in hospital B) and recovered on average significantly more LNs than pathologists did (18.5 vs 12.2 in hospital A and 16.6 vs 13.2 in hospital B). PAs harvested a significantly higher percentage of LNs <5 mm than pathologists did (64.2% vs 53.7%). The percentages of colon cancer patients eligible for adjuvant chemotherapy due to inadequate LN sampling alone were significantly higher for cases dissected by pathologists than for those dissected by PAs (17.3% vs 1.1% in hospital A and 13.1% vs 3.4% in hospital B)
Conclusions PAs contribute to patient safety since they recover more and, in particular, smaller LNs from colorectal resection specimens than pathologists do. Moreover, they help to reduce costs and morbidity by reducing the number of patients eligible for adjuvant chemotherapy due to inadequate LN sampling alone.
In the past decade, patient safety and error reduction in surgical pathology have increasingly received attention. Several authorities developed quality standards and guidelines that contribute to ensuring patient safety. Regarding the pathologist, several factors contribute to errors, including distractions and workload increase.1 The authors of a Turkish study proposed to decrease pathologists’ workload by deploying ‘pathologists’ assistants’ (PAs).2
There is a shortage of pathologists in the Netherlands. Therefore, starting from 2009, our institution has been training PAs to take over certain routine activities in order to save pathologists’ valuable time, which they can use instead to focus on microscopy and on interdisciplinary meetings. Based on our own experience and on results of other studies, we expect that adequately trained PAs can contribute to patient safety.3 ,4
For the scope of this study, we focused on lymph node (LN) harvest from colorectal resection specimens, which is a labour-intensive and time-consuming activity. Due to the nature of their job, PAs can usually spend more time focusing on harvesting LNs than pathologists can. This is the main reason why we expect PAs, on average, to harvest more LNs per specimen than pathologists do.
Metastases in LNs form an important prognostic factor in determining eligibility for adjuvant chemotherapy,5 as 5-year survival considerably decreases when ≥1 LNs are malignant. Also, the number of harvested LNs has been shown to be closely related to recurrence and survival.6–11 In a study on 480 node-negative patients, 5-year survival increased from 51% when <10 LNs were sampled to 69% when 10–19 LNs were harvested and rose to 71% when >19 LNs were examined.12
Nonetheless, 20–40% of patients with presumed negative LN status eventually die as a result of their cancer.13 A significant number of particularly small (metastatic) LNs might be missed during grossing, as 45–78% of metastatic LNs measure <5 mm in diameter.14–21 Precise examination and retrieval of as many LNs as possible is therefore of paramount importance for reducing the risk of understaging.
The recommended minimum number of LNs varied from 6 to 17 in previous studies.8 ,9 ,13 ,22–30 The Dutch guideline31 recommends a minimum of 10 LNs, whereas several international guidelines recommend a minimum of 12 LNs.32 ,33 According to the seventh edition of the American AJCC Cancer Staging Manual, a minimum of 10–14 LNs must be examined.34 In the Netherlands, in 2007 only 65% of all stage I and stage II cancer cases had adequate LN sampling (≥10 LNs evaluated).35 The Dutch Surgical Colorectal Audit (DSCA) of 2009 showed that in 73% of colon cancer cases and in 58% of rectal cancer cases, ≥10 LNs were evaluated. These percentages rose to 83% and 68%, respectively, in 2011.36 At our institution, we also aim at retrieving more LNs per specimen. Previous studies show percentages of ≥12 LNs per rectal cancer case varying between 5% and 64% after neo-adjuvant chemo-radiotherapy and between 55% and 88% without prior chemo-radiotherapy.37–43
Colon cancer patients (≥T2 stage) receive adjuvant chemotherapy if they present with ≥1 metastatic LNs or if they are at a high risk of developing recurrences. Patients at a high risk present with T4 tumours, poor differentiation, bowel obstruction, lymphovascular invasion, perineural invasion, bowel wall perforation or an insufficient number of sampled LNs.
The purpose of this study was to evaluate the role of PAs in the improvement of patient safety by comparing their LN harvest results from colorectal resection specimens with those of pathologists. Moreover, we evaluated whether PA deployment could theoretically result in a reduced proportion of colon cancer patients who could have been eligible for adjuvant chemotherapy due to inadequate LN sampling alone.
This study included colorectal cancer patients in two Dutch hospitals (A and B), whose pathology services were provided by Symbiant. We retrospectively reviewed 557 pathology reports of patients who were surgically treated for colorectal cancer between January 2008 and November 2011 (table 1). The surgical technique did not change during the study period and was the same for every surgeon. In each hospital, ten pathologists and seven PAs were responsible for harvesting LNs.
The PAs in this study were officially trained by Symbiant and the Centrum Bioscience en Diagnostiek, Hogeschool Leiden. The PAs started their training by observing a pathologist processing resection specimens. They then learned to dissect specimens and sample LNs themselves, under direct supervision of a pathologist. After proving their skills to the pathologist, PAs began working independently, with the pathologist acting as a back-up. In 2009, PAs had taken over from pathologists in 19.8% of cases. By 2010, they sampled LNs in 53.2% of the cases and by 2011, 66.7% of all specimens were grossed by PAs.
Resection specimen processing
Colon and rectal resection specimens were randomly assigned to a pathologist or a PA, based on their schedules, and processed routinely. After fixation in neutral buffered formaldehyde, the adipose tissue was sectioned at 1–2 mm intervals to gather the LNs, and the number of LNs per cassette was accurately counted and registered.
The independent samples t-test or ANOVA was performed to compare the average number of harvested LNs per specimen between the subgroups. χ2 analysis was used to compare the percentages of cases with ≥10 LNs evaluated, the percentages of LNs <5 mm dissected by PAs and pathologists and the percentages of patients eligible for adjuvant chemotherapy due to inadequate LN sampling alone.
We performed logistic regression (OR, 95% CI and p value) for univariate (UV) and multivariate (MV) analysis. Factors that contributed significantly to adequate LN sampling in the UV analysis were considered for MV analysis. p Values <0.05 were considered as statistically significant. All p values reported are two-sided.
LN harvest by pathologists and PAs
Table 2 shows the percentages of colorectal cancer cases which were adequately sampled by pathologists and PAs as well as the average numbers of harvested LNs per specimen. Compared with pathologists, PAs more often succeeded in sampling ≥10 LNs per specimen (83.2% vs 60.9% in hospital A, p<0.0001; and 79.2% vs 67.6% in hospital B, p=0.019). Not only did PAs more often harvest ≥10 LNs per specimen than pathologists did, their average LN count per specimen also proved higher (18.5 vs 12.2 and 16.6 vs 13.2 in hospitals A and B, respectively, with p<0.0001 for both hospitals).
The role of tumour location in LN harvest
Table 3 shows the relation between tumour site and LN sampling. The percentage of cases in which ≥10 LNs were sampled was significantly lower in rectal resection specimens than in right colon resection specimens (p<0.0001 for both hospitals). Also, the average number of LNs per specimen harvested in rectal resection specimens was significantly lower than in resection specimens from the right colon (p<0.0001 for both hospitals), and in hospital A it was also lower than in resection specimens from the left colon (p=0.003).
LN harvest in the rectal cancer subgroup
As rectal cancer cases seem to be among the most difficult to sample adequately, we evaluated this subgroup separately (table 4). In hospital A, PAs retrieved ≥10 LN in 63.2% of cases and pathologists harvested ≥10 LNs in 38.2% of cases (p=0.035), whereas in hospital B the percentage achieved by PAs (61.0%) was almost equal to that of pathologists (62.5%).
Neo-adjuvant chemotherapy had no significant effect on the percentage of adequately sampled cases (46.7% vs 61.0% with and without neo-adjuvant chemotherapy, respectively, p=0.197).
Factors contributing to adequate LN harvest
UV analyses showed that the following factors contributed significantly to adequate LN harvest (based on OR): tumour site, tumour diameter, T-stage and PA. MV analysis showed that the factors tumour site, tumour diameter and PA still contributed to the harvest of ≥10 LNs (based on adjusted OR). After correction for other factors, PA was associated with a greater effect on sampling ≥10 LNs (OR 2.671; 95% CI 1.726 to 4.135) than was shown in UV analysis (OR 2.318; 95% CI 1.578 to 3.406) (table 5).
Recovery of LNs<5 mm
Hospital A counted 113 patients with ≥1 metastatic LNs. Comparison of the maximum diameters of all LNs from these patients showed that PAs had recovered a significantly higher percentage of LNs <5 mm than pathologists had recovered (table 6).
Eligibility for adjuvant chemotherapy due to inadequate LN sampling
Table 7 shows the percentages of colon cancer patients who could have been eligible for adjuvant chemotherapy based solely on inadequate LN harvest. In the subgroup of patients dissected by PAs, significantly fewer patients were eligible for adjuvant chemotherapy due to inadequate LN harvest alone than in the subgroup of patients dissected by pathologists (1.1% vs 17.3% in hospital A, p<0.0001; and 3.4% vs 13.1% in hospital B, p=0.007).
The aim of this study was to determine the value of PAs for patient safety by comparing LN harvest results from colorectal resection specimens between PAs and pathologists.
Sampling LNs is time-consuming and particularly difficult to achieve in rectal cancer.
Our study showed a significant improvement in LN harvest after deployment of PAs in two Dutch hospitals. We evaluated the average number of harvested LNs per specimen and the percentage of cases with ≥10 LNs examined, which in the Netherlands is considered to be the minimum number needed to assess prognosis and the indication for adjuvant chemotherapy. Multivariate analysis confirmed that the PAs were responsible for improved LN yield. Results were comparable when the AJCC threshold of 12 LNs was used (data not shown). Reasons for these results were not researched but are likely to be varied and to include practical circumstances. The main reason for the use of PAs is to optimise the use of pathologists’ time.44 PAs are generally dedicated to this type of task, whereas pathologists’ increasing workloads can lead to time constraints. Due to the nature of their job, PAs can usually spend more time, with fewer distractions, searching for LNs than pathologists can spend.44
Results of the 2011 DSCA showed that ≥10 LNs had been evaluated in 83% of colon cancer cases included in the audit. The PAs working in the two hospitals in which this study was performed found ≥10 LNs in 92.0% and 88.1% of colon cases, whereas the pathologists found ≥10 LNs in 70.4% and 70.0% of colon cases. When considering the more widely used recommendation of ≥12 LNs, the respective success rates do not change (≥12 LNs in 87.4% and 77.1% of colon cases by PAs, and in 58.0% and 50.0% by pathologists). We should add, however, that pathologists and PAs had the Dutch guideline of ≥10 LNs in mind.31
Rectal cancer patients generally receive neo-adjuvant radiotherapy and/or chemotherapy, resulting in tumour downstaging.45 Neo-adjuvant therapy, however, induces LN shrinkage.46 Added to the fact that LNs in the rectal mesentery are already inherently smaller, as is the mesentery itself (compared to colonic mesentery), it becomes increasingly difficult to meet the criterion of examining ≥10 LNs in rectal resection specimens.47 ,48 The problem of insufficient LN harvesting is therefore particularly relevant for rectal cancer. In our study, all metastatic LNs measured <5 mm in 16.7% (4/24) of the rectal cancer cases with ≥1 metastatic LNs. A previous study showed that all metastatic LNs measured <5 mm in 32% of 98 rectal cancer cases with metastatic LNs, and in 8% of these cases all metastatic LNs measured <2 mm.48 Nonetheless, PAs in hospital A significantly contributed to harvesting ≥10 LNs (63.2% compared to 38.2% by pathologists) in the subgroup of rectal cancer cases in the present study.
This study concurs with comparable studies.3 ,4 Reese et al reported a significant increase in LN harvest when a single PA was made responsible for LN harvest in a single centre.3 The average numbers of harvested LNs increased, from pathologists averaging 13.6 LNs per specimen to the PA averaging 19.7 LNs. In terms of percentages, pathologists adequately sampled 58.4% of cases (≥12 LNs evaluated), while the PAs correctly sampled 84.0% of the cases. Galvis et al4 studied a single centre employing two PAs, who retrieved more negative LNs from axillary dissection and colorectal resection specimens than pathology residents did. As these studies analysed the work of just one or two PAs, it cannot be deduced that PAs in general perform better than pathologists. However, our two-centre study, which measured results of seven PAs per hospital, corroborated the validity of the findings of these previous studies.
Our study is the first to show a direct link between PA deployment and improved harvest of small (<5 mm) LNs (64.2% vs 53.7% by PAs and pathologists, respectively). In the final year of our study, by which time PAs were responsible for 66.7% of cases, 46.2% of all metastatic LNs measured <5 mm, which is consistent with previous studies (45–78%).14 ,16–20
Negative LN status (after adequate LN sampling) can spare colon cancer patients adjuvant chemotherapy. Theoretically, in hospital A, the percentages of colon cancer patients who were eligible for adjuvant chemotherapy because of inadequate sampling alone could have been reduced from 17.3% with pathologists harvesting LNs to 1.1% with PAs harvesting LNs. In hospital B this average percentage would have fallen from 13.1% to 3.4%.
We demonstrate that PA deployment significantly improves patient safety and leaves pathologists more valuable time to spend on tasks which require their specific expertise. Another reason to employ PAs can be to cut costs, as pathologists are relatively expensive.44 Future studies will need to examine the time and cost effectiveness of PAs, focusing on turnaround times and resubmission rates.
In conclusion, PAs contribute to patient safety since they recover more and, in particular smaller LNs from colorectal resection specimens than do pathologists. Moreover, they help to reduce costs and morbidity, by reducing the number of patients eligible for adjuvant chemotherapy due to inadequate LN sampling alone.
Adequate evaluation of at least 10 lymph nodes (LNs) is needed to establish negative LN status in colorectal cancer patients.
Especially small lymph nodes (LNs) should be evaluated as a significant percentage of LNs <5 mm (45–78%) contain metastases.
Colon cancer patients might be eligible for adjuvant chemotherapy solely based on inadequate LN sampling.
Pathologists’ assistants significantly contribute to patient safety, because their work directly results in improved harvest of (small) LNs, which would theoretically result in a decline in the number of patients eligible for adjuvant chemotherapy.
Acknowledgements The authors would like to thank Alice van Berne and Lisette van Hulst for their excellent editing work and Tjeerd van der Ploeg for his statistical assistance.
Contributors CK, WS and NJ designed the study; CK and NJ interpreted and reviewed the results, and drafted the manuscript; WS and GM provided clinical insight in the data; MA and AS were responsible for grossing; PvD provided statistical advice; HvS, WS, GM, MA, AS, VC, PvD and NJ reviewed the manuscript; NJ takes responsibility for the paper as a whole.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.