icon-folder.gif   Conference Reports for NATAP  
 
  50th ICAAC
Boston, MA
September 12-15, 2010
Back grey_arrow_rt.gif
 
 
 
Clinical, Virological and Immunological Characteristics of Patients With Discordant Phenotypic and Genotypic (Ultra-Deep Sequencing) Tropism Test Results: Analysis of Tropism Calls in MOTIVATE and 1029
 
 
  Reported by Jules Levin
ICAAC 2010 Sept 14 Boston
 
D Chapman1, H Valdez1, M Lewis2,
I James2, RA McGovern3, LC Swenson3,
PR Harrigan3, J Heera4
 
1Pfizer Inc, New York, NY, USA; 2Pfizer Global Research and Development, Sandwich, UK; 3BC Centre for Excellence in HIV/AIDS, Vancouver, BC, Canada; 4Pfizer PGRD New London, CT, USA

Harrigan PR, et al. 5th IAS 2009. Abs WELBA101; 2. Swenson L, et al. 12th EACS 2009. Abs.PE3.3/2.

1. Swenson L, et al. 12th EACS 2009. Abstract PE3.3/2; 2. Swenson L, et al. 47th IDSA 2009. Abstract 297

The majority of the discordant results shown in the orange slice were R5 by Trofile and were reclassified as non-R5 by 454.
 
This is similar to the percentage of patients who were reclassified as having non-R5 virus by the enhanced sensitivity Trofile assay in the TN MERIT study. There were also 15 patients shown in the yellow slice who were R5 by 454 and non-R5 by Trofile.

* Includes Germany, France, Spain, Netherlands, Poland, Sweden, Switzerland, Italy and Belgium, UK

aIncludes 7 patients with X4 virus by Trofile.
This graph splits out Trofile R5 results by strata of percent non-R5 virus detected by 454 shown along the x axis.
 
As expected, concordant Trofile R5 results were associated with <2% non-R5 virus by 454.
 
The majority of samples with discordant Trofile R5 and 454 non-R5 results had between 2-30% non-R5 virus by deep sequencing, while only ~5% had >30% non-R5 virus.
 
Most (631/784; 80%) Trofile R5 results were associated with <1% non-R5 virus The highest proportion of non-R5 Trofile results (38/105; 36%) was seen at levels above 60% CXCR4 use Approximately one third (20/58; 34.5%) of samples with non-R5 virus present at levels above 60% returned an R5 Trofile result
 
Approximately one third (20/58; 34.5%) of samples with non-R5 virus present at levels above 60% returned an R5 Trofile result

Now shown in the red bars on this slide are patients with Trofile non-R5 results The highest proportion of concordant non-R5 Trofile results were seen at levels above 30% non-R5 virus by 454.
 
All patients with pure X4 virus by Trofile fell into this group with most displaying >98% non-R5 virus.
 
There were a small number of patients with Trofile non-R5 results who had discordant 454 R5 results
 
Most (631/784; 80%) Trofile R5 results were associated with <1% non-R5 virus The highest proportion of non-R5 Trofile results (38/105; 36%) was seen at levels above 60% CXCR4 use Approximately one third (20/58; 34.5%) of samples with non-R5 virus present at levels above 60% returned an R5 Trofile result
 
Approximately one third (20/58; 34.5%) of samples with non-R5 virus present at levels above 60% returned an R5 Trofile result

This slide is a graphical comparison of population sequencing and Trofile results using 454 tropism calls as the reference standard.
 
Population sequencing results are shown in the striped bars and Trofile results are shown in the solid bars.
 
Shown on the left is a high concordance of Trofile with Population sequencing in predicting 454 R5 calls.
 
(Build)
When 454 was used as the reference standard for non-R5 tropism calls, population sequencing detected approximately 13% more non-R5 virus than the original Trofile test.
 
Not depicted on this slide is the fact that population sequencing was highly concordant with Trofile and 454, with the majority of 80% having concordant results across all three tests.
 

These graphs quantify the amount of non-R5 virus detected by deep sequencing in terms of absolute log counts of non-R5 virus shown on the left and percentages of non-R5 virus shown on the right.
 
Discordant groups are represented by the yellow and orange bars shown in the center of each graph Patients with concordant non-R5 results (shown in the red bars) had the highest proportion of non-R5 virus with a median of 47% non-R5 virus Patients with discordant results which were R5 by Trofile and non-R5 by 454 (shown in the orange bars) also had a high median absolute non-R5 viral load comparable to that of patients with concordant non-R5 virus, although the percent of non-R5 virus in this group was much lower.
 
This suggests that for Trofile calls, the percent of non-R5 virus is more important than the absolute non-R5 VL.

This graph shows virologic responses in terms of log change from baseline through 24 weeks of treatment.
 
Patients with the highest non-R5 VLs shown in the orange and red lines were all non-R5 by 454 and had similarly weaker responses irrespective of their Trofile results.
 
Patients with concordant R5 results had the strongest virologic responses (shown by the green line) The small number of patients with discordant results of R5 by 454 but non-R5 by Trofile (shown in the yellow line) had better virologic responses than those with non-R5 454 results.
 
These patients had slightly worse virologic responses than those with concordant R5 results, despite percent X4 virus < 0.5% suggesting that non-R5 viral load is more important to predict antiviral responses Due to the small number of pts in this group, we plan to confirm this finding in subsequent data.