Sir,
In the interesting paper on utility of HPV genotype triage in the primary HPV cervical screening using HCII method, authors of the article (1) claimed that there existed an additional advantage of HPV genotype triage in combination of HPV 16/18/52/58 in the detection high grade CIN/HSIL on Histology. Their false interpretation of their data had resulted in incorrect conclusions for their findings on large cervical HPV screening data (n= 5456) in the study. The others comments are about i) duplicating information in tables 1 and 2, tables 3 and 4, and ii) Plots of ROC curves requires to be based on all viral load measurements instead of 3 points plotted by authors in their article.
For the purpose of explaining the controversy in interpretation, we used data given in the article (1) to provide a correct comparative picture of various triaging approaches (Table 1). Authors in their study arrived at conclusion that additional genotyping for HPV52/58 to HPV16/18 was beneficial in triaging women with positive HPV test using HCII in a primary cervical screening. The interpretation of authors was false as they were looking at the benefit of only increase in sensitivities and ignored reducing specificities in providing conclusions to their study. They could have avoided ending with a conflicting message. For an acceptable screening tool, a very high specificity and a reasonably good detection capacity (sensitivity) is needed while minimising referral rate for colposcopy directed biopsies. A correct comparative picture from re-tabulation of data published by authors in their article is presented in Table 1. This is apparent from Table 1 that for only different genotypes or different genotype combinations with cytology, sensitivity elevates up to 96.4% while specificity reduce to below 50%. On the other hand, Cytology triaging had balanced sensitivity (64.3%) and specificity (82.1%) along with lowest referral rate (2.2%) for colposcopy in comparison with other strategies. In addition, we computed likelihood ratio for positive (LR+) and negative (LR-) tests using the data presented by authors and shown in Table 1. For example, LR+ 3.6 for cyto triage means one in 28 with positive test is with disease. LR+ 1.8 for HPV 16/18/52/58 triage means one in 54 with positive test is with disease.
This shows that triaging with HPV genotype with 16/18/52/58 is no better than cytology triage alone. Hence the conclusion drawn by authors were incorrect and requires to be modified or corrected as the study in fact supports the cytology triage in comparison to HPV genotype 16/18 or 16/18/52/58 triaging.

Table 1 Comparison of Triage strategies of cytology and genotype after HCII primary cervical screening
Triage Method Colpo Required (%) Sensitivity Specificity OR LR+ LR-
Only cyto +ve 2.2 64.3 82.1 8.2 3.6 0.43
Only genotype
HPV16/18 3.6 28.3 83.3 1.9 1.7 0.86
HPV 16/18/52/58 3.5 85.7 51.8 6.4 1.8 0.28
HPV 16/18/31/58 3.6 91.1 47.0 90.0 1.7 0.19
Combination Cyto + Genotype
HPV 16/18 2.8 71.4 70.5 5.9 2.4 0.41
HPV 16/18/52/58 3.6 94.6 44.6 14.2 1.7 0.12
HPV 16/18/31/58 3.7 96.4 40.6 18.5 1.6 0.09
Data source: Surapan K etal (2016) (1)
Likelihood ratios :LR +ve and LR-ve were computed by us for each of the triage approach.
References:
1. Khunamornpong S, Settakorn J, Sukpan K, Suprasert P, Srisomboon J, Intaraphet S, et al. (2016) Genotyping for Human Papillomavirus (HPV) 16/18/52/58 Has a Higher Performance than HPV16/18 Genotyping in Triaging Women with Positive High-risk HPV Test in Northern Thailand. PLoS ONE 11(6): e0158184. doi:10.1371/journal.pone.0158184

Dr Smita Asthana MD
Scientist D
Dr Satyanarayana Labani, Ph D, FSMS*
Scientist G & Head
Division of Epidemiology & Biostatistics,
National Institute of cancer Prevention & Research (NICPR), (ICMR)
Ministry of Health & Family Welfare, Govt of India
Sector-39, Noida, UP, India
*corresponding to satyanarayanalabani@yahoo.com