N-nitroso-dimethylamine, or NDMA, is one of the specific nitrosamine that I think is most likely to be cause for concern. NDMA has been found in notable quantities in cured meats and other similar food products. This study found concentrations of about NDMA highest in sausages at about 10 micrograms per 100 grams in sausage. This same study reported bacon as containing about .5 micrograms per 100 grams, but when cooked this number seems to increase.

A study by Richard Peto et al. measured different intakes of NDMA on the health of rats and found that NDMA increased liver cancer. Regarding dose they stated the following; "Since the experiment continued into extreme old age, effects became clearly measurable even at a dose of only 0.01 mg/kg per day". So even .01 mg per kg of body weight increased cancer. And for those doubting the validity of a rat study on humans, look at this:

This graph, taken from this paper, suggests the potential carcinogenic effects of NDMA (via formation of DNA adducts) is similar in rats and humans. Yes, humans are still not rats, but this should make the extrapolation from a rat study slightly more reasonable.

Moving on, let's imagine a real world scenario; if a 65 kg person were to consume a pound of sausages in a day that might give them about 45 micrograms, or .045 mg of NDMA. This would amount would be less for a pound of bacon. The dose of NDMA that still seems to increase liver cancer (via that rat study) for that 65 kg person is .65 mg.

So eating a pound of sausages every day might give you 7% of the dose of NDMA shown to increase liver cancer incidence in rats. Eating a pound of bacon every day would give you much less, probably no more than about 4% of the dose. This is pretty low. So I'm now unconvinced that NDMA makes a few slices of bacon or some pieces of cured meat all that risky. And I really haven't seen much to suggest that other nitrosamines in food are going to increase cancer either.

Anyway, I think this is a good question and hope to hear more people weigh in.

Edit:

My answer is already lengthy, but since this subject has received some attention lately I thought it was due time I updated my answer with additional info.

First, I wanted to briefly point out two issues with using with using rodents to try and determine carcinogenic thresholds for nitrosamines. One the one hand, rodents don't live as long as humans and thus are less able to model long term low dose exposure. On the other hand, rodent diets are typically refined and the likelihood that they contain non-essential phytochemicals is slim to nil. This is important because, at least in rats, garlic, green tea, tomatoes, ellagic acid, and quercetin, among other compounds, have been shown to inhibit the carcinogenic potential of specific nitrosamines. Also, again in rats, the tumor promoting effect of N-nitrosomethylurea (and I would bet most nitrosamines) seems to increase directly with increased linoleic acid intake from vegetable oils.

Ok, moving on for now. I really wanted to find studies measuring the effect of eating cured (nitrate/ite treated) meat compared to uncured meats on cancer incidence with no additional variables. So far, no luck in human or animal studies. Still, I found some interesting studies worth looking at:

1. In a composite of several studies, groups of 100 total rats were injected with azoxymethane (a chemical that causes colon cancer) and fed either beef, chicken, or bacon for 100 days (this procedure was repeated using higher and lower fat diets). The researchers then looked for aberrant crypt foci (ACF) throughout the rat's colons. ACF are used as crude markers of tumor promotion. The bacon diet not only didn't appear worse, but in certain cases actually seemed to result in better ACF measurements.

2. Another study fed 9 mice diets either composed of 18% hot dogs or not for 7 days. At the end of this period there was no difference in 7-methyldeoxyguanosine levels in the two groups. 7-MedG isn't itself a significant promoter of cancer, but is measured because it can be a crude marker for the formation of other oncogenic DNA adducts. Weak study, but interesting.

3. In two related studies, actual humans (not rodents!) were, in the first study, fed either fresh red meat or a vegetarian diet. In the second study subjects were fed either processed meat or a vegetarian diet. The trials lasted 14 days. Indicators of colon cancer risk were examined, including intestinal cell DNA breaks,and indicators of oxidized pyrimidines and altered purines. Comparing the fresh red meat and processed meat diet was suggestive of a slightly worse effect of the processed meat, although interestingly enough the vegetarian diet group produced significantly worse values on all things measured compared to both meat groups.

So based on all of this, let me expand and reiterate my original opinion a bit (and finish this increasingly long answer): I think that some gently cooked meat cured with nitrates/nitrites (with all other things being equal) in a diet rich in plants and low in omega-6 rich seed oils is unlikely to be a significant contributer to cancer risk.