Trial Information

For Aim 1, two 5ml pre-treatment blood and corresponding tumor samples will be obtained from
NSCLC patients at the National University Health System. Sampling will be organized to avoid
the blood samples being the first sample taken after skin puncture to minimize contamination
with skin epithelial cells. For the first blood sample, CTCs will be isolated and retrieved
using the NUS developed CTC bio-chip according to methods described previously.21 From
second blood sample, CTCs will be isolated, fixed on the chip and stained for EpCAM, CD45
and DAPI to assess for cell purity and quantity. DNA will be extracted from the retrieved
CTCs and tumour samples, and analyzed exon 19 deletion, L858R and T790M mutations by digital
PCR on the Fluidigm Biomark according to methods described previously.23 EGFR mutation
status in blood and tumour samples will then be compared for their concordance.

For Aim 2, patients with NSCLC being treated with gefitnib will be approached. Two 5ml blood
samples will be obtained pre-treatment (baseline) and then every 4 weeks of treatment (one
cycle of gefitinib) from NSCLC patients at the National University Health System.

CTCs will also be obtained and analysed from patients on another protocol receiving
gefitinib/ hydroxychloroquine. (B/08/196. A phase II with a lead in phase I study to examine
the tolerability, safety profile and efficacy of Hydroxychloroquine and Gefitinib in
advanced Non-Small Cell Lung Cancer.) In this study, CTCs are already being collected. Hence
the investigators intend to use samples from B/08/196 for CTC analysis using the CTC biochip
platform.

For each timepoint, CTCs will be isolated, retrieved and analyzed for EGFR mutation status
as described above. Clinical response will be determined using the RECIST criteria.34
Associations between pre-treatment EGFR mutation type quantities and best clinical response
will be assessed by Fisher's exact test. Association between trends in EGFR mutation type
quantities and tumour size over all treatment cycles will be assessed by observation as
performed in the study by Maheswaran et al.9

The investigators plan to analyze 30 patients over one year for both aims of this study.
Some cases will have relevant samples (serial blood samples and corresponding tumour) for
both aims. The sample size is based on that of the study by Maheshwaran et al.9 and also the
likely volume of suitable subjects during 1 year at NUHS. Relevant data for sample size
estimation is otherwise lacking. In particular, the frequency of T790M mutations in a
relevant patient population is lacking, highlighting the lack of adequate analytical systems
for its assessment such as the one proposed in this study.