Correlation between pre-operative diagnosis and post-operative pathology reading in pediatric neck masses--a review of 281 cases.

OBJECTIVE

Describe the pre-operative diagnosis correlation of pediatric neck masses with the final post-operative pathology reading. Evaluate if added imaging studies were associated with higher frequency of correct diagnosis.

STUDY DESIGN

Retrospective case series review.

SETTING

Tertiary Academic Medical Center.

SUBJECTS AND METHODS

Data was collected from 281 pediatric patients with an undiagnosed neck mass that underwent open biopsy from January 1986 to December 2011. Data collection included pre-operative accuracy and relative contributions of clinical clues and imaging studies.

RESULTS

Pre-operative and post-operative diagnoses match varied for each category of diagnoses. The difference in distribution of correct pre-operative diagnoses between 6 categories of neck masses was statistically significant. The highest number of cases with correct pre-op diagnosis was seen with congenital masses which were correctly diagnosed with 75% of cases (n=109), followed by benign tumors which were diagnosed with 73.7% of cases (n=19). When CT scans were included in the pre-op work up, it was associated with a non-significant trend toward a less frequent correlation between pre and post-operative diagnosis in congenital, nodal inflammatory and miscellaneous masses. However, with regards to the diagnosis of benign tumors, CT scan was associated with a trend toward higher proportion of correct pre-op diagnosis. In cases where US was included in the evaluation, we found a trend toward less frequent correlation with post op diagnosis in benign tumors, nodal inflammatory, non-nodal inflammatory and miscellaneous diagnoses and an increase in accuracy for congenital masses.

CONCLUSION

Certain types of pediatric neck masses are easy to diagnose likely due to their classic presentation. Failure to diagnose masses often occurs when the clinical picture is vague or non-specific. Ancillary imaging studies do not always correlate with increased accuracy of diagnosis, particularly when the clinical clues are typical.