BOSTON ( TheStreet) -- This week's Biotech Stock Mailbag fisks the recently published short thesis on Exact Sciences.

The backstory: On Wednesday, "Alpha Exposure" -- the pseudonym of a hedge fund -- published its short thesis of Exact Sciences on Seeking Alpha. The article -- Exact Sciences -- The Rotten Sniff Test -- is a very skeptical look at the company's Cologuard colon-cancer screening test. "Alpha Exposure" accuses Exact Sciences of hiding important data from Cologuard's pivotal phase III study and cherry-picking data that has been released in order to trick investors into believing the non-invasive colon cancer test is more accurate than it really is.

"Alpha Exposure" said it plans to publish four more articles about Exact Sciences on Seeking Alpha but as of Thursday mid-day, just the single article had been posted.

On June 14, I described the bull thesis on Exact Sciences, as explained to me by a different hedge fund investor who is long the stock. And yes, he also requested anonymity.

Here's what my hedge fund manager source had to say Thursday about the Exact Sciences' short thesis explained by "Alpha Exposure" in Wednesday's article:

"Well, it's just amazing that this guy is using the wrong numbers. His whole thesis is that Exact Sciences is cherry-picking data, but in fact, that's exactly what he did."

Two hedge funds battling over the future of Exact Sciences and its colon-cancer screening test! It's always good to know both sides of the story, but who's right? Based on what I've heard and read so far, I lean towards the bull thesis because I believe "Alpha Exposure" is actually being more manipulative with data than anything the fund accuses Exact Sciences of doing.

Let me explain. Two main points:

First, "Alpha Exposure" accuses Exact Sciences of misleading investors by setting up an easy and rigged comparison between Cologuard sensitivity and specificity and historical data on FIT sensitivity and specificity. The fund points to Exact Sciences' recent investor slides comparing Cologuard to older, previously published FIT data as proof the company is misrepresenting the accuracy of its screening test.

"Alpha Exposure" is wrong. In the pivotal "DEEP-C" study conducted by Exact Sciences, all patients underwent screening for colon cancer using Cologuard and FIT. Here's what Exact Sciences CEO Kevin Conroy said about the data comparing Cologuard to FIT. The quote comes from the company's May 1 conference call:

Let's turn now to how Cologuard performed against FIT, fecal immunochemical test, during the DeeP-C trial. Cologuard achieved superiority to FIT, not just for pre-cancer sensitivity but for cancer sensitivity as well. Our Cologuard trial was designed with the sequential approach to analyzing the performance of Cologuard compared to FIT. After determining that Cologuard achieved non-inferiority to FIT for cancer sensitivity, our statisticians were able to make a calculation regarding superiority. We are very pleased to report that Cologuard demonstrated superiority to FIT for cancer sensitivity with a statistically significant P value. Claims of superiority to FIT for both cancer and pre-cancer detection will be included in our FDA submission.

Exact Sciences has not yet disclosed the detailed data showing the numerical superiority of Cologuard over FIT because the results are included in a manuscript of the study submitted for publication in a medical journal.

Again, Conroy, this time from the company's July 23 conference call:

"...We have prepared our analysis of the trial's full data set in collaboration with Principal Investigator and submitted a manuscript in June to a premier peer reviewed medical journal. We hope to have the paper published before the end of the year."

This explains why Exact Sciences uses historical data on FIT in its investor slides.

What we know already from the DEEP-C study is Cologuard's sensitivity to cancer: 92%, pre-cancer sensitivity: 42%; and specificity: 87%. We also know these Cologuard numbers are superior to FIT, we just don't know by how much until the study is published.

Second, "Alpha Exposure" contends that even when Exact Sciences discloses the Cologuard vs. FIT comparison from the DEEP-C study, the result will be bogus because the company chose a FIT comparator that is obsolete and pre-destined to deliver subpar results.

If Exact Sciences was interested in an honest comparison between Cologuard and FIT, the company would have used the InSure FIT as the comparison screen in the DEEP-C study, "Alpha Exposure" says.

InSure FIT is an easier-to-use, more accurate version of the old FIT, with cancer sensitivity of 88%, pre-cancer sensitivity of 43% and specificity of 97%. Across all three measures, InSure FIT would beat Cologuard, "Alpha Exposure" claims.

As proof, the hedge fund points to a November 2006 study published in Cancer comparing InSure FIT to an older fecal occult blood test.

Here's where "Alpha Exposure" cherry picks data to unfairly bolster its Exact Sciences short thesis. If you read the abstracted results from the Cancer paper only, indeed, InSure Fit's sensitivity and specificity appear just as described by "Alpha Exposure" in its Seeking Alpha article.

But if you read the text of the Cancer study, you'll get to this important caveat in the discussion:

The true-positive rates observed in this study for those found to have cancer in either population are likely to overestimate the true sensitivity of screen-detected cancers in the general population because the diagnostic cohort was included. The design is powerful for determining relative sensitivity, as results are paired for these tests. Better estimates of true sensitivity come from the screening population (Tables 3 and 4), although because not everyone had a colonoscopy, results shown in Table 4 represent the upper limit of actual sensitivity within the limitation of the number of detected cases."

In other words, InSure Fit's 88% sensitivity for cancer and 43% sensitivity for pre-cancer are not realistic, real-world measures. They're over-estimations, according to the study authors.

What does a more realistic evaluation of Insure Fit's accuracy in detecting cancer and pre-cancer look like? Thankfully, the Cancer paper tells us:

An estimate of true sensitivity can be obtained from results in those screening cases where colonoscopy was performed (Table 4). Although numbers were relatively small, InSure was significantly more sensitive for cancer (75% vs. 37.5%) and significant adenomas (27% vs. 15%).

Oh! When the study authors looked only at patients where colonoscopy was performed -- the most accurate method to diagnose colon cancer -- the performance metrics of Insure Fit are downgraded significantly.

Cancer sensitivity is really 75%, not 88%. Pre-cancer sensitivity is 27%, not 43%, as "Alpha Exposure" claims.

The hedge fund doesn't include these more real-world Insure Fit numbers in its Seeking Alpha article because they're numerically inferior to Cologuard -- and that just ruins the entire premise of the short thesis.

The remaining four parts of the "Alpha Exposure" short thesis on Exact Sciences might be more convincing. Part one was not.

-- Reported by Adam Feuerstein in Boston.

Adam Feuerstein writes regularly for TheStreet. In keeping with company editorial policy, he doesn't own or short individual stocks, although he owns stock in TheStreet. He also doesn't invest in hedge funds or other private investment partnerships. Feuerstein appreciates your feedback; click here to send him an email.