I have a 13 shot pano, each in vertical format. So, mea culpa, I had some sensor spots - well quite a lot actually. When I did this a couple of years ago, I fixed them all on one frame and Sync'd them from the 'master' to the other 12, and did the pano in PhotoShop and printed it 6 feet wide. Worked fine.
Using those same images with LR CC, it would appear that the images are taken into the Merge without applying the sensor spot corrections.
I re-listend to JK's tutorial and note that some adjustments are not applied to the input images (are in fact discarded) with the recommendation they be applied to the output image.
However, in the case of sensor spot corrections - seems they should be left corrected on input. In this case - 10 corrections per image - means roughly 100+ in the final with additional complications as the images blend.
An oversight - or is it intention that sensor spot corrections are discarded.
Or is there a buried option to retain the corrections on input