Detection of micrometastases in sentinel lymph nodes from melanoma patients: direct comparison of multimarker molecular and immunopathological methods.

The technique of sentinel lymph node (SLN) dissection is a reliable predictor of metastatic disease in the lymphatic basin draining the primary melanoma. Reverse transcription-polymerase chain reaction (RT-PCR) is emerging as a highly sensitive technique to detect micrometastases in SLNs, but its specificity has been questioned. A prospective SLN study in melanoma patients was undertaken to compare in detail immunopathological versus molecular detection methods. Sentinel lymphadenectomy was performed on 57 patients, with a total of 71 SLNs analysed. SLNs were cut in slices, which were alternatively subjected to parallel multimarker analysis by microscopy (haematoxylin and eosin and immunohistochemistry for HMB-45, S100, tyrosinase and Melan-A/MART-1) and RT-PCR (for tyrosinase and Melan-A/MART-1). Metastases were detected by both methods in 23% of the SLNs (28% of the patients). The combined use of Melan-A/MART-1 and tyrosinase amplification increased the sensitivity of PCR detection of microscopically proven micrometastases. Of the 55 immunopathologically negative SLNs, 25 were found to be positive on RT-PCR. Notably, eight of these SLNs contained naevi, all of which were positive for tyrosinase and/or Melan-A/MART-1, as detected at both mRNA and protein level. The remaining 41% of the SLNs were negative on both immunohistochemistry and RT-PCR. Analysis of a series of adjacent non-SLNs by RT-PCR confirmed the concept of orderly progression of metastasis. Clinical follow-up showed disease recurrence in 12% of the RT-PCR-positive immunopathology-negative SLNs, indicating that even an extensive immunohistochemical analysis may underestimate the presence of micrometastases. However, molecular analyses, albeit more sensitive, need to be further improved in order to attain acceptable specificity before they can be applied diagnostically.

Modeling sequencing errors by combining Hidden Markov models.

Among the largest resources for biological sequence data is the large amount of expressed sequence tags (ESTs) available in public and proprietary databases. ESTs provide information on transcripts but for technical reasons they often contain sequencing errors. Therefore, when analyzing EST sequences computationally, such errors must be taken into account. Earlier attempts to model error prone coding regions have shown good performance in detecting and predicting these while correcting sequencing errors using codon usage frequencies. In the research presented here, we improve the detection of translation start and stop sites by integrating a more complex mRNA model with codon usage bias based error correction into one hidden Markov model (HMM), thus generalizing this error correction approach to more complex HMMs. We show that our method maintains the performance in detecting coding sequences.